Hydrocarbon Processing 2019-03

HYDROCARBON PROCESSING
MARCH 2009
MARCH 2009
HPIMPACT SPECIALREPORT BONUSREPORT
Top initiatives INSTRUMENTS GAS PROCESSING

in automation AND NETWORKS DEVELOPMENTS
INSTRUMENTS AND NETWORKS
Nobel Laureate Wireless, soft sensors, New methods

new DOE head OPC and H2 detection treat natural gas
www.HydrocarbonProcessing.com
Our focus on quality produces RENTECH
boilers tough enough for any specs.
This built-in engineering and production muscle will save you time and costs in
both installation and long term maintenance. Why not get boilers that are tough
enough to always make you look good? Take our factory tour and see for yourself
(and while in Abilene, we’ll treat you to the best steak you’ve ever eaten!).
Fired Package Boilers / Wasteheat Boilers / Heat Recovery Steam Generators

www.rentechboilers.com Maintenance & Service Strategies / Boiler Repair Services / SCR and CO Systems
Select 64 at www.HydrocarbonProcessing.com/RS
MARCH 2009 • VOL. 88 NO. 3
SPECIAL REPORT: INSTRUMENTS AND NETWORKS
29 Wireless networks improve refinery operation

Smart instruments and secure wireless communications enable enhanced operations
and asset management
Cover Illustration courtesy of Emerson
G. Martin Process Management. See related article,
“Wireless networks improve refinery
33 OPC UA: an end user’s perspective
The updated specification relies on Web services for its data transportation providing
operation,” page 29.
significant advantages
R. Kondor
39 Soft sensor modeling using artificial neural networks

Here are guidelines for proper construction
V. Nandakumar
45 Hydrogen gas detection

Combining detection systems improves safety
E. Naranjo
HPIMPACT
17 Networking, alarm
BONUS REPORT: GAS PROCESSING DEVELOPMENTS management, security
among top initiatives
48 Fine-tuning demercaptanization process: A case study
Optimizing caustic concentrations and reactor temperatures improved 19 Coke drum delivery
acidic compound removal without installing new equipment marks project milestone
Z. Mallaki and F. Farhadi at Texas refinery
55 What are the opportunities to construct liquefaction

facilities at the Arctic Circle?
19 Pace of economic
decline forecast to slow
in first half of 2009
Building and operating natural gas plants in the high latitudes pose numerous challenges
D. A. Wood and S. Mokhatab 19 Nobel Laureate Chu
selected to head US
59 In-line laboratory and real-time quality management
An in-depth look at NIR spectroscopy
Department of Energy
M. Valleur
ROTATING EQUIPMENT/RELIABILITY
66 Auxiliary pumps and support systems

for process machinery
Proper system design and operation are critical to plant uptime and reliability
J. R. Brennan
PROCESS DEVELOPMENTS COLUMNS

69 Consider practical conditions for vacuum unit modeling
A good simulation model is a tool that reveals critical operating conditions
9 HPIN RELIABILITY
Unreliability, global
and can be applied to daily operations procurement and you
R. Yahyaabadi
11 HPIN EUROPE
Sacrificed to the money
OPERATOR TRAINING/MANAGEMENT
system: engineering
workforce
77 From dynamic ‘mysterious’ control to dynamic
‘manageable’ control 13 HPINTEGRATION
Instructional design strategies and delivery methods for bridging the DMC chasm STRATEGIES
S. M. Ranade and E. Torres A good alarm
management strategy
DEPARTMENTS 86 HPIN CONTROL

CDU overhead double-
7 HPIN BRIEF • 15 HPIN ASSOCIATIONS • 17 HPIMPACT • drum configuration
21 HPINNOVATIONS • 25 HPIN CONSTRUCTION • 82 HPI MARKETPLACE •
85 ADVERTISER INDEX
View this month’s LETTERS TO THE EDITOR online at:
Houston Office: 2 Greenway Plaza, Suite 1020, Houston, Texas, 77046 USA
Mailing Address: P. O. Box 2608, Houston, Texas 77252-2608, USA
Phone: +1 (713) 529-4301, Fax: +1 (713) 520-4433
E-mail: [email protected]
London Office: Nestor House, Playhouse Yard London, EC4V 5EX, UK,
Phone: +44 (0) 20 7779 8800, Fax: +44 (0) 20 7779 8996/8899
Publisher Mark Peters [email protected]
EDITORIAL
Editor Les A. Kane
Senior Process Editor Stephany Romanow
Managing Editor Wendy Weirauch
Process Editor Tricia Crossey
Reliability/Equipment Editor Heinz P. Bloch
News Editor Billy Thinnes
European Editor Tim Lloyd Wright
Contributing Editor Loraine A. Huchler
Contributing Editor William M. Goble
Contributing Editor Y. Zak Friedman
Contributing Editor ARC Advisory Group (various)
MAGAZINE PRODUCTION
Director—Editorial Production Sheryl Stone
Manager—Editorial Production Chris Valdez
Artist/Illustrator David Weeks
Manager—Advertising Production Cheryl Willis
ADVERTISING SALES
See Sales Offices page 84.
CIRCULATION +1 (713) 520-4440
Director—Circulation Linda K. Johnson
SUBSCRIPTIONS
Subscription price (includes both print and digital versions): United
States and Canada, one year $140, two years $230, three years $315.
Outside USA and Canada, one year $195, two years $340, three
years $460, digital format one year $140. Airmail rate outside North
America $175 additional a year. Single copies $25, prepaid.
Because Hydrocarbon Processing is edited specifically to be of greatest
value to people working in this specialized business, subscriptions are
restricted to those engaged in the hydrocarbon processing industry, or
service and supply company personnel connected thereto.
Hydrocarbon Processing is indexed by Applied Science & Technology
Index, by Chemical Abstracts and by Engineering Index Inc. Microfilm
copies available through University Microfilms, International, Ann
Arbor, Mich. The full text of Hydrocarbon Processing is also available
in electronic versions of the Business Periodicals Index.
ARTICLE REPRINTS
If you would like to have a recent article reprinted for an upcoming
conference or for use as a marketing tool, contact us for a price quote.
Articles are reprinted on quality stock with advertisements removed;
options are available for covers and turnaround times. Our minimum
order is a quantity of 100.
For more information about article reprints, call Cheryl Willis at
+1 (713) 525-4633 or e-mail [email protected]
HYDROCARBON PROCESSING (ISSN 0018-8190) is published monthly by Gulf Publishing Co., 2 Greenway Plaza,
Suite 1020, Houston, Texas 77046. Periodicals postage paid at Houston, Texas, and at additional mailing office.
POSTMASTER: Send address changes to Hydrocarbon Processing, P.O. Box 2608, Houston, Texas 77252.
Copyright © 2009 by Gulf Publishing Co. All rights reserved.
Permission is granted by the copyright owner to libraries and others registered with the Copyright Clearance Center
(CCC) to photocopy any articles herein for the base fee of $3 per copy per page. Payment should be sent directly to
the CCC, 21 Congress St., Salem, Mass. 01970. Copying for other than personal or internal reference use without
express permission is prohibited. Requests for special permission or bulk orders should be addressed to the Editor.
ISSN 0018-8190/01.
GULF PUBLISHING COMPANY

John D. Meador, President/CEO
Mark Peters, Vice President
Ron Higgins, Vice President
Maggie Seeliger, Vice President
Pamela Harvey, Business Finance Manager
Part of Euromoney Institutional Investor PLC.
Other energy group titles include:
World Oil®
Petroleum Economist
Publication Agreement Number 40034765
Printed in U.S.A
䉳 Select 151 at www.HydrocarbonProcessing.com/RS
I N T E N S E H E AT. A G G R E S S I V E C H E M I C A L S . E X T R E M E C O L D .
A HIGHER DEGREE
OF PERFORMANCE.
982°C TO BE EXACT.
MATERIAL TECHNOLOGY
FOR HYDROCARBON
PROCESSING
THERMICULITE®
835 Spiral Wound Filler
• Handles the toughest applications
• Outperforms graphite and fiber
• Provides total freedom from oxidation
• Offers true outage-to-outage assurance
• Reduces inventory requirements
ALSO AVAILABLE IN:

• 815 Tanged Sheet Your Global Gasket Provider
• 815 Cut Gaskets
• 845 Flexpro ™ (kammprofile) Facing log onto:
www.flexitallic.com
or call:
Thermiculite ®
US +1 281.604.2400
Innovative. Versatile. Complete. UK +44 (0) 1274 851273
HOW WOULD YOU RATHER ACCESS REMOTE
GAS LINES TO MEASURE MOISTURE CONTENT?
ON FOOT. ONLINE.
New AMETEK 5100 NCM™ laser analyzer

combines integrated moisture verification
with Ethernet-based user interface.
AMETEK’s dependable 5100 NCM noncontact moisture analyzer for natural
gas applications has all the convenience, performance and features you
demand. Moisture reading verification, combined with its Ethernet/Web
browser-based interface, eliminates your need to be on-site at all! The 5100
NCM features all-digital signal processing and an accuracy to ±4 ppm over
a 5-2500 ppm range, with a 0.25 lb./MMscf limit of detection. It meets
CL1 DIV 2 Groups A-D approvals*.
With simple analyzer setup and system checks, the 5100 NCM provides
readout information anywhere it’s needed, reliably and online—no complex
software required. Remote readouts and diagnostics lower maintenance
costs and reduce downtime. With no exposed components and an
IP-65/NEMA 4 weatherproof enclosure designed to endure -20°C to
+50°C, it’s rugged as all outdoors.
So rest your feet, and leave the rest to AMETEK. Learn more at:
412-828-9040 or www.ametekpi.com
*Other approvals pending. Select
Select123
58 at
at www.HydrocarbonProcessing.com/RS
www.HydrocarbonProcessing.com/RS
HPIN BRIEF
WENDY WEIRAUCH, MANAGING EDITOR
[email protected]
Report monitors Canadian oil sands projects. The recent unprecedented ■ Multinational oil
shifts in crude oil’s price and the weakening global economy is impacting smaller companies
proposing oil sands projects. “When we couple the weak economy and volatile price of oil perspectives
with continued rising costs for oil sands operators, the margins for greenfield producers are
shrinking,” says a new study from the Canadian Energy Research Institute (www.cera.ca). There is a “renewed need to react” to
Margins for producers are being absorbed by continued cost increases, much of which is supplying global demand when world-
due to professional and skilled labor, materials and equipment, and greenhouse gas emis- wide economies pick up, said Jesus
sions costs. Under present economic conditions, global oil prices need to be closer to C$90 Reyes Heroles, director general of
WTI to support new proto-typical oil sands projects over the next 30 years, according to Pemex. He presented his views at the
this analysis. CERAWeek conference, held recently
in Houston. Pemex is committed to
North American LNG imports are set to rise, according to one recent analysis. increasing Mexico’s refining capacity
In light of recent history, and the longer term outlook for growth in domestic US shale and avoiding engaging in “stop and
gas, many industry commentators and analysts are suggesting that the outlook for LNG go” behavior on project investments.
imports into North America is bleak. “However, while it is fair to say that regas capacity
has undoubtedly been overbuilt, Wood Mackenzie believes that the medium-term outlook Mr. Heroles said that his company is
for LNG in North America is not as dire as other commentators are suggesting,” says a searching for new “modalities” to coop-
company study (www.woodmacresearch.com). The new forecast projects growth for LNG erate with other national and interna-
imports into North America from 2009 to 2014. Wood Mackenzie predicts that the tional oil companies. He also stressed
medium-term outlook for LNG in North America is that LNG imports will increase from the urgency in retaining valuable human
1.7 Bcfd in 2009 to 4.2 Bcfd in 2014. resources so as to counter the past few
years’ critical workforce shortages.
How will new US administration influence energy stocks? Analysts
with Casey Research have examined potential policies that Washington could implement Jiping Zhou, vice president of China
and how these might affect a particular industry sector. “A bull market will come for the National Petroleum Corp. and president
traditional energies in the long run; the problem lies in the shorter term, in the instability of of PetroChina Co. Ltd., gave his perspec-
America’s energy portfolio,” says this investment viewpoint. The coal industry could be in tive on the state of the industry to the
for a hard time under President Obama. His proposed tough 100% cap-and-trade system conference attendees. He noted that
will make coal plants uneconomical to run. “As natural gas is already one of the cheapest the long-term fundamentals for product
power technologies available, the industry would weather a cap-and-trade system better supply and demand have not changed
than coal,” according to this research. by the present global slowdown. Calling
this a “temporary difficult time,” he
US demand for specialty additives used in gasoline and other fuels projected an upswing in his country’s
is forecast to increase 2.9%/yr to $1.3 billion in 2012. Above-average growth for deposit economic activity in late 2009. His com-
control agents—the largest segment of fuel additives—will continue to support the mar- pany intends to maintain its “moderate
ket, according to a new study from The Freedonia Group, Inc. Regulations are forecast to increase” of industry investment.
boost demand for cold-flow improvers, which are necessary to increase the performance
of ULSD and biodiesel in colder climates. Corrosion inhibitors are also expected to show Tony Hayward, group chief executive of
steady growth through 2012 as these additives are needed to counteract the effects of higher BP, in his address, stressed the importance
oxygenate levels in fuel. Corrosion inhibitors and additives used in diesel fuel, such as cold- of looking through the here and now to
flow improvers, will show the fastest growth, says this report. the longer term of improved economic
activity and, consequently, heavier glob-
Maintaining capital project competitiveness in a slow economy. Over al oil and product demand. “The future
the past three to four years, the engineering and construction industry has struggled with is not canceled,” despite present dreary
how to get a massive number of complex domestic and international projects completed business headlines, he affirmed. His com-
safely, on time and within budget while providing quality deliverables. The single most pany’s business strategists are operating
influential negative aspect of projects during this time (as defined by benchmarking from under the “important reality” that 80%
CII, IPA, ECC and others) was the lack of skilled resources at all levels—within both the of the world’s energy will be coming
owner and contractor organizations. “Many economic forecasts indicate that the capital from fossil fuels in 2030.
project industry will be down for approximately two to three years and then jump to levels
similar to 2006–2008,” according to Stephen L. Cabano, president of Pathfinder LLC, a Mr. Hayward supports a cap-and-trade
project management consultancy. He cautions that the industry would be best served by system for lowering emissions, and also
investing in training and mentorship to ensure that project teams have the skill sets and emphasized the importance of a step-
tools for addressing the challenges of 2010 and beyond. HP change in energy R&D investments. HP
HYDROCARBON PROCESSING MARCH 2009

I7
build on our foundation
Our depth of knowledge and experience gives UOP

customers a head start.
As the global leader in technology solutions for the petroleum

refinery industry since 1914, UOP understands what it takes to
help our customers achieve and sustain success. Today, with the
support of our new parent company, Honeywell, we reaffirm our
commitment to leadership in customer satisfaction and
innovation. From equipment design and consulting to process
technology and products like high-performance catalysts and adsorbents, UOP is the one
global company that can consistently add value to your project.
Process Technology • Catalysts • Adsorbents • Performance Equipment • Profitability Consulting

UOP LLC, 25 East Algonquin Road, Des Plaines, IL 60017-5017, USA phone: +1-847-391-2000 fax: +1-847-391-2253 www.uop.com
©2007 UOP LLC. All Rights Reserved.
HPIN RELIABILITY
HEINZ P. BLOCH, RELIABILITY/EQUIPMENT EDITOR
[email protected]
Unreliability, global procurement and you

Allow us to suggest that you engage in a “reality check” on the bearings and seals) in global procurement, here’s why you should
subject of equipment unreliability, global procurement and your brace for potentially very serious trouble.
own role in the matter. Please examine, realistically and objec- The dimensional and material property-related accuracy of
tively, the direction in which much of industry seems headed. spare parts that have an impact on the plant’s safety and reliability
Then, take action if the danger signs we bring to your attention must comply with rigorous specifications and quality control.
pertain to you. Therefore, start by identifying the approximately 5 to 7% of parts
To begin with, we hope that your contributions to the safety, and components in your critical machinery that have such reli-
profitability and sound utilization of the employer’s (or share- ability impact and assume your manager will be pleased with your
holder’s) assets are highly valued. However, if you have decided doing this identifying. Next, take tangible remedial steps. Alert
or are being asked to keep your reliability concerns to yourself, it others to the urgency of only consenting to global purchasing of
may be time to readjust your thinking. We believe a true reliability these parts after appending or invoking rigorous specifications
professional must let others know about valid concerns and must and quality control.
then take discrete steps to have these reliability and uptime issues Unless proven otherwise, you should assume that the lowest bid-
properly addressed and resolved. der utilizes neither quality control nor exacting specifications. Per-
Suppose you are consistently making solid contributions and haps this explains why it is the lowest bidder. You must provide and
these are neither valued nor acknowledged. In that case you might sometimes personally write a specification for these critical parts.
consider updating your resume and seek work at a location where Once critical spare parts (even the ones originating from vendors
experienced reliability engineers are in demand. On the other hand, accepting your specifications and professing to have quality control)
start with an honest appraisal of the real value of your own contribu- are delivered to your facility, the job is far from finished. You must
tions. Acknowledge that there is room for improvement with every add value by personally verifying the full specification compliance
human being. Are you having a positive influence on others? Are of these parts. Alternatively, take responsibility by arranging for
you really adding value to the enterprise every step of the way? competent inspectors that verify specification compliance of the
For example, it would make little sense if you were to confine critical spare parts received. These parts should be accepted by the
your contribution to telling management that you’re “concerned” storeroom clerk only after compliance has been verified. The clerk
that synthetic lubricants might be incompatible with certain paints, can then proceed to tag and preserve the parts for future use.
or if you merely challenged the recommendation that synthetics
should be put into your cooling tower gearboxes. If you were to Understand your role and carry out your duties.
voice similar “concerns” on about two-dozen other peripheral The role of a true reliability professional has been spelled out
issues you will have added no value and will have nudged your in many books and articles. A professional is not just “a pair of
employer closer to becoming a second-tier, low-profitability com- hands.” The ones that have become top contributors in their area
pany. Instead, follow up on your concerns and establish whether of expertise participate in reliability audits, engage in structured
or not these are justified. Along these lines, and as an example, it root-cause failure analyses that culminate in eliminating repeat
should take you no more than 10 minutes to ascertain that the failures, develop repair specifications and condemnation limits,
synthetic gear oil under consideration really only attacks acrylic i.e., parameters beyond which parts can no longer be repaired,
house paint, and that your gearbox interiors are painted with a assemble work processes and procedures to match best-of-class
highly stable epoxy paint not prone to those attacks. In research- competition, perform life cycle cost analyses and propose training
ing the matter, you might uncover that most of your competitors plans for themselves and future reliability engineers. It would seem
have, for decades, used one of the synthetics being considered, and logical that reliability professionals become familiar with how
that their cooling tower fan gears have accrued an average life of 20 their best-of-class colleagues function in these roles and have been
years. So, understand the life cycle cost implications and become able to keep their jobs in good times and in bad times. HP
an advocate of change instead of a skeptic voicing unspecified or
vague concerns passed down by word-of-mouth. LITERATURE CITED
You may contact the author for a list of references.
Living with global procurement. If your company is
presently involved in global procurement of critical machine
components, take note of a few very important facts and draw
The author is the Equipment/Reliability Editor of HP. A practicing engineer and
the right conclusions. Global procurement often implies buying ASME Life Fellow with close to 50 years of industrial experience, he advises process
from the lowest bidder or from parties that offer rapid delivery. If plants on maintenance cost-reduction and reliability upgrade issues. His 16th and
your company favors this simple version of a global procurement 17th textbooks on reliability improvement subjects were published by John Wiley &
Sons in 2006.
approach and includes certain OEM parts (such as compressor
I9
What Has Thomas Russell Co. Been Doing?
We Have Built 30 n Standard 40, 60, 120 & n Fractionation
Plants Processing 200 MMSCFD or Custom
Designed Cryogenic Plants
n Crude Topping
3 BCFD in n Gas or Liquid Amine
n Reformers
4 Years. Treaters up to 1000 + GPM
n Refrigeration Units for
LPG Recovery
DESIGN • BUILD • INSTALL
7050 South Yale, Ste. 210

Tulsa, Oklahoma 74136
Phone: 918/481-5682 Fax: 918/481-7427
www.thomasrussellco.com
HPIN EUROPE
TIM LLOYD WRIGHT, EUROPEAN EDITOR
[email protected]
Sacrificed to the money system: engineering workforce

Because of the shortage of an artificial commodity known as they get for what they sell is falling. T: Why?
money, people who produce a real commodity known as oil will PT: Companies and consumers usually borrow money to buy,
shortly be losing their livelihoods and, quite possibly, their homes. build or make new things—and that uses energy—but now they
We’re used to this cycle, but does it really have to be this way? can’t. T: Why?
Those of us who worked in or near the oil industry through the PT: Well, the banks aren’t lending money like they used to.
1990s already have the scent of what’s coming. Mergers, consoli- T: Why?
dation, cost-savings and canceled projects all mean that any time PT: Too many people or companies are defaulting on loans
soon job cuts are due in a corridor near you. they made in the past. In a modern economy, the way to supply
ConocoPhillips, the first of many perhaps, has announced that money for repaying loans and the interest is through the writing
it is cutting 4% of its overall workforce, slashing capital spending of new loans. T: Why?
by 18% and writing off $34 billion in assets because of falling PT: Well, when banks write loans, the government allows them
energy prices. So, there it is, at least 1,300 job cuts on the table for to use that promise of the borrower to repay to create new money
starters. We’re just in that part of the economic cycle. at that point. In a process that the economist J.K. Galbraith
As my six-year-old daughter Thalia would say: “Why?” She has described as “so simple the mind is repelled,” that’s where money
a charming, although on occasions, somewhat testing way with the comes from. It enters the money supply of the nation, formerly
word “why.” It is simply inserted at the end of each presumed answer as privately issued paper derivatives of the assets in the bank’s safe.
until, if the interviewee is willing, the conversation turns to matters These are known as private bank checks, but today the credit of
of principle or the nature of things more deep and fundamental than the bank is legally interchangeable by the bank with the fiat cur-
ice cream or why a third viewing of Tom and Jerry is not okay. rency of the nation… the pounds, dollars or pennies we use to buy
things. The borrower repays the bank and must pay interest to the
Father and daughter discourse. Alas, that such intellec- lender, but that creates a shortage in the money supply. T: Why?
tual rigor isn’t more common in the adult. Just why people in the PT: The private bank checks—today just numbers typed into the
energy industry are losing jobs is a question well worth asking. borrower’s bank account—are created and convertible to ordinary
After all, does the world no longer need energy? Are engineers currency, but the interest is not created. That means the amount of
and chemists, geologists, project managers and the supporting loans issued and fiat currency created must always grow. T: Why?
infrastructure not performing a function as critical today as last PT: Without more and more borrowing, there won’t be enough
summer when a metric ton of heating oil cost in excess of $1,000, money generally available for the repayment of the interest on
and the stuff it was made from famously hit $147/bbl? the loans. A growing proportion of the borrowers, represented in
In spite of all the warnings heard about security of supply, are economics by the formula I /(P+I ), will be foreclosed by the bank,
we really so sure of ourselves that we can begin to dismantle the transferring their assets to the bank. T: Why?
infrastructure for providing it? Of course, the knee-jerk answer is PT: Well, to cater for the repayment of the interest, there must
“It’s the market, stupid,” but I think we need to scratch deeper. be continuous, exponential growth in the economy so that new
The new head of the UK’s Financial Services Authority, Lord loans are taken and money is created. Recently, this borrowing
Turner, seems to think that we all should apply some of young has had to be undertaken by governments. But if the number of
Thalia’s rigor. He says: “Across the world, there has been an intel- foreclosures reaches a certain point, people get in a panic and stop
lectual failure to understand that we were building a system which issuing loans altogether. T: Why?
has huge systemic risks.” PT: Banks know what economist Irving Fisher knew, that banks
I propose using Thalia’s infinitely recurring why and a dialogue don’t lend money; they, in fact, lend “promises to supply money
between father and child for the rest of this month’s column. I’m they do not possess.” If this promise looks like it may not be met
not saying the father has all the right answers, but in common because wholesale lending is founded on bad loans, then that’s a
with many of us, he’s put in some study since the banking system problem and the system can come crashing down, leading to a
collapse began. situation where there is no longer enough money to facilitate the
Pappa Tim: I can’t come up and cut paper shapes with you essential functions of society—including developing and providing
right now. T: Why? energy resources. And that’s why people are losing their jobs. HP
PT: I’m writing an article about people losing their jobs in the
oil industry. T: Why? The author is HP’s European Editor and has been active as a reporter and
PT: Well, the oil companies don’t have enough money any conference chair in the European downstream industry since 1997, before which
more to pay them their wages. T: Why? he was a feature writer and reporter for the UK broadsheet press and BBC radio.
Mr. Wright lives in Sweden and is founder of a local climate and sustainability
PT: Well, the companies and the consumers who are their initiative.
customers don’t have as much money as before, and so the price
I 11
Shouldn’t you know your safety
loops from the inside out?
With a smart safety instrumented system you can. DeltaV SIS. Smart.
DeltaV SIS is the only logic solver to digitally diagnose and automatically proof test elements of your entire safety loop. No
more second guessing–they’re all in view. And, it provides this valuable information seamlessly back to the operators
without a lot of extra/difficult to maintain integration effort–ensuring your process shuts down when it should and not
when it shouldn’t.
DeltaV SIS is the first to use digital intelligence and diagnostics to enable safer facilities, improve availability, lower life-cycle
costs, and ease regulatory compliance. For more insight, visit: EasyDeltaV.com/SIS
TM
The Emerson logo is a trademark and a service mark of Emerson Electric Co.©2009 Emerson Electric Company
HPINTEGRATION STRATEGIES
LARRY O’BRIEN, CONTRIBUTING EDITOR
[email protected]
A good alarm management strategy

The ISA S18.02 standard provides a much needed, standard- tion of alarm management solutions will provide more metrics, offer
ized framework for implementing an effective and sustainable improved identification of alarm floods and provide easier hooks to
alarm management strategy in refineries, petrochemical plants metrics that will allow users to access the data they need.
and other process plants. Alarm manage- ISA S18.02 outlines best practices for
ment continues to be a serious issue for ■ Once it has been finalized, alarm strategy development for both new
process automation end users. According and existing facilities. ISA S18.02 covers
to NIST, an average of $20 billion is lost this standard has the potential all aspects of alarm strategy development,
in the US manufacturing industry every from alarm philosophy to rationaliza-
year due to abnormal conditions. Forty to greatly reduce the number tion, detailed design, implementation,
percent of these incidents can be directly operation, maintenance, management of
attributable to human error. When you
of incidents in process plants change, monitoring and assessment, and
consider that alarm systems are the criti- and will have a major impact auditing. The standard also builds on the
cal point between emerging abnormal sit- fine work already done by the Abnor-
uations and the operator action required on unplanned downtime and mal Situation Management Consortium
to alleviate those situations, it becomes (ASM), the Engineering Equipment and
obvious that a refinery’s alarm manage- profitability. Materials Users Association (EEMUA)
ment strategy can have a huge impact on and NAMUR. To date, the EEMUA has
throughput and profitability. had the closest thing to a best-practices document that can address
common issues surrounding today’s alarm systems. In fact, there
The state of process alarm management. To date, was a formal liaison between NAMUR and the EEMUA commit-
there has been little in the way of standards activities in the area tees when establishing the S18.02 standard.
of alarm management. Certain groups, such as EEMUA and
NAMUR, have outlined best practices for alarm management, but State of the standard. The ISA S18.02 standard is very
there have been no formal standards development activities. You close to becoming finalized. The most recent ballot results at the
may ask, “Why is a standard even needed?” It’s needed because October 31st meetings showed that 74% of responding members
the overall state of the process alarming strategy at most owner/ approved the standard and it has been provisionally approved by
operator companies is shabby at best. There’s no cost associated the committee, pending incorporating comments. The final stan-
with adding alarms to today’s DCSs. As a result, end users are dard could be available by the end of the summer this year.
swamped with alarms, only some of which require any real action
to be taken. Many operators have reached the point where they Adopting ISA S18.02 to measure customer methods.
spend a disproportionate amount of time dealing with alarms. ARC anticipates that regulatory bodies, the insurance industry
The situation is only going to get worse as alarms and alerts start and other health, safety and environment-related concerns, such
coming in from plant asset management systems, intelligent field as HSE in the UK or OSHA in the US, will adopt ISA S18.02 as
devices, fieldbus-based safety systems and so on. a basis for examining customer practices in alarm management as
they relate to overall process safety and sustainability. These orga-
What is ISA S18.02? The ISA S18.02 standards development nizations have not yet had a standard against which to measure
activity provides owner/operators and other end users with a company performance in alarm management. Don’t be surprised
blueprint for developing an effective alarm management strategy. if your insurer comes into your plant and asks how you are man-
Once it has been finalized, this standard has the potential to aging your alarms according to the ISA S18.02 standard so your
greatly reduce the number of incidents in process plants and will operators are not getting flooded with alarms. HP
have a major impact on unplanned downtime and profitability.
ISA S18.02 is directed at people who use control systems and pre-
scribes a life cycle-based approach to managing alarms. It guides
end users through the whole process of establishing a life cycle Larry O’Brienis ispart
The author partofofthe
theautomation
automationconsulting
consultingteamteamat at ARC
ARC covering
covering the
the
process
processindustries,
industries,andandananHPHPcontributing
contributing editor. HeHe
editor. is responsible
is responsible for for
tracking the
tracking
program where alarms are set up and rationalized in a consistent
market for process
the market automation
for process systems (PASs)
automation systems and(PASs)
has authored
and hasthe PAS market
authored the stud-
PAS
way and reviewed for effectiveness. ies for ARC
market sincefor
studies 1998.
ARCMr. O’Brien
since hasMr.
1998. alsoO’Brien
authored hasmany
also other
authoredmarket
many research,
other
ISA S18.02 does not tell automation suppliers how to design their strategy
market and custom
research, research
strategy andreports on topics
custom researchincluding
reportsprocess
on topicsfieldbus, collaborative
including process
alarm systems, but it does help them make modifications to their partnerships, total automation
fieldbus, collaborative market trends
partnerships, and others. He
total automation has been
market with
trends andARC since
others.
January
He has1993,
been and
withstarted his career
ARC since with 1993,
January marketandresearch
startedin the
his field instrumentation
career with market
alarm management solutions that will allow end users to put together
markets.
research in the field instrumentation markets.
their own alarm management program or strategy. The next genera-
I 13
Spray Spray Spray Spray
Nozzles Control Analysis Fabrication
Spray Injector Solutions

Improve Performance, Extend Service
Life and Reduce Maintenance
We have dozens of ways to help optimize the performance of your spray injectors,
quills and spool pieces. Here are just a few:
U Assistance with nozzle selection and injector placement in the gas stream –
critical factors to application success
U Validation using 3D modeling capabilities and spray testing in our labs
based on your operating conditions ensure performance goals are met
U Recirculating, air- or liquid-cooled, multiple nozzle designs and more
to meet any quality standard or extreme engineering requirement
U Retractable, flexible and multi-directional designs are available
to minimize maintenance and service interruptions
Learn More at spray.com/injectors

Visit our web site for helpful literature on key considerations in
spray injector design and guidelines for optimizing performance.
Our solutions include injectors for: Computational Fluid Dynamics

(CFD) is often used to help fine
U Distillation columns tune injector performance
requirements and placement
U Regenerator bypass
U FCCU water wash
U Fractionator water wash
U Pollution control equipment
U Steam quench
U And more
In the US and Canada: 1-800-95-SPRAY | 1-630-665-5000 | spray.com | [email protected]

HPIN ASSOCIATIONS
BILLY THINNES, NEWS EDITOR
[email protected]
Association news in brief

2009 Industrial Automation facets to any successful gathering—a learn- The website from which much of the
Safety and Security ing component and networking oppor- outreach is managed is www.energyto-
Symposium tunities. Following the panel discussion, morrow.org. Ms. Ganz said the outreach
The 2009 Industrial Automation Safety there will be a dinner for attendees at the efforts can be considered a success. After
and Security Symposium will take place Casa Rio restaurant on the Riverwalk. evaluating the tone of coverage and level of
April 22–23 at the Marriott Houston Hobby engagement, she thinks the media, public
Airport in Houston, Texas. This event is pro- Houston BMA luncheon and lawmakers were forced to reconsider
duced by the International Society of Auto- At the Houston Business Marketing some opinions. For instance, in June 2007,
mation (ISA). The symposium will address Association (BMA) luncheon in January, the tone of monitored media stories and
technical and business issues associated with three speakers delved into educational blog postings was 2–1 against the energy
identifying and mitigating safety hazards in and perceptional outreach efforts from industry. By August 2008, this tone was
industrial environments. Additionally, this the energy industry to students, educators, flipped, with coverage 2–1 in favor.
year’s symposium will include additional members of the media, legislators and the Tommy Lyles, a communications man-
technical focus on cyber security threats to general public. ager at Chevron, concluded the program
industrial environments and design con- Bill Pike spoke first, as a representative by speaking about a game his company
siderations engineers must consider when for the Society of Petroleum Engineers had developed with an eye toward educat-
designing industrial processes and safety (SPE). He discussed the SPE’s educational ing middle school and high school stu-
instrumented systems. The symposium will website, www.energy4me.org. He then dents about energy policy. Called “Wel-
provide an in-depth look at today’s safety described other outreach efforts, including come to Energyville,” the game can be
technologies and procedures. The event an energy education kit for K–12 class- accessed by visiting www.willyoujoinus.
is intended to create a forum where paper rooms and an oil and natural gas book com/energyville.
presentations and panel discussions transfer for students. According to Mr. Pike, SPE
information from the leaders and experts on distributed 6,500 books in 2008 and plans SPAR conference to take
safety and control to industry professionals. on translating the book into multiple lan- over Denver
Technical theme areas include: safety guages in 2009. SPAR’s 2009 conference convenes
instrumented systems, alarm management, Susan Ganz, an American Petroleum March 30-April 1in Denver, Colorado.
industrial security and lowering cost of Institute (API) member and marketing The focus of the conference includes 3D
capital and return on investment through executive for Schlumberger, was next on laser scanning, mobile surveying, asset
safety and security projects. To register, the program. Her remarks were about management, CAD/GIS integration and
visit www.isa.org/safetysymposium. API’s education strategy. According to security planning. Charles Matta, director
her, a survey from August 2007 rated the of federal buildings and modernization
GPA convention seeking energy industry 20th out of 21 industries for the General Services Administration
young professionals in serving customers. With that in mind, (GSA), will give a keynote presentation
The 88th annual Gas Processors Asso- API developed an e-advocacy goal of on the GSA’s use of 3D scanning for its
ciation (GPA) convention takes place bringing more balanced media coverage of BIM initiatives. The Shaw Group’s Andy
March 8–11 in San Antonio, Texas. Any the industry while also raising energy lit- Guard will offer a case study on how his
midstream young professionals that will eracy levels. One element of this approach firm is using laser scanning for industrial
be at the convention are encouraged to was founding a communications center to plant applications. On the education side,
participate in an event called “Fueling tell the industry’s story, with capabilities of there is much talk about the 3D laser scan-
Your Future.” The event features a special rapid response to correct inaccurate infor- ning boot camp, which will be delivered
discussion with John Gibson, CEO of mation. Outreach by company CEOs was by SPAR’s advisory board.
ONEOK. Following Mr. Gibson’s remarks also encouraged and chats were arranged New exhibitors at SPAR 2009 include:
will be a panel of industry experts ready to with influential audiences. ClearEdge3D, CSA, IXSEA Land and Air,
field questions about career opportunities Ms. Ganz said the specifics of the strat- TechSoft 3D and Velodyne. There are also
and options. The panelists are expected to egy involved 120 events in 55 markets. several association sponsors, including
be long time veterans of the gas process- These events included keynotes, panels and the ASTM, the American Society of Civil
ing industry and should have the ability to a partnership with Newsweek that sometimes Engineers, CyArk, the International Asso-
answer any questions proffered, no mat- utilized “influencer salons.” She was also ciation of Forensic and Security Metrol-
ter how technical or far-fetched. The GPA proud of a touring interactive technology ogy and the Society of Piping Engineers
believes this event will combine two crucial exhibit that has visited 20 state capitals. and Designers. HP

I 15
When the
right reaction
matters ...
Trust BASF Refining Catalysts

At crucial moments, the right reaction matters. When you
are looking for the right reaction from your refining catalyst,
turn to BASF. Our technical experts will recommend the right
catalyst from our innovative product line that will achieve the
desired reaction. The end results will be more of the products
that you want. When the catalyst is right, the reaction will be
right. Trust BASF.
䡵 FCC Catalysts 䡵 FCC Additives 䡵 FCC Solutions
For more information, please visit
www.catalysts.basf.com/refining
HPIMPACT
WENDY WEIRAUCH, MANAGING EDITOR
[email protected]
Networking, alarm Bayway refinery in Linden, New Jersey, Looking through the crystal ball, respon-
management, security commented that the refinery wanted all dents do not see the recession lasting; they
departments to be on the same page. “We said that the biggest business challenge over
among top initiatives are looking at advanced controls projects the next five years will be workforce develop-
ISA recently conducted an online sur- to integrate more of the refinery’s units ment, followed closely by aging out of work-
vey to find out what automation industry together,” Mr. Mitchell says. ers and profitability concerns. “Baby Boom-
observers and practitioners felt that near- Others simply just want to understand ers” leaving the industry remains an issue.
term trends were going to be. what their equipment is telling them.
When survey participants were asked “We need to move into OPC to get more Outlook in Europe. In economic terms,
which technology their facility would rely on data,” according to Robert Dusza, project the 2009 outlook for the European control
for 2009, the top choice was networking at and tech support manager at Manchester and instrumentation sector seems slumping,
21%. “With wireless being the rage through- Water and Sewer in Manchester, Connecti- with layoffs and project cancellations becom-
out the industry, you would think it would cut. “Since we buy from the lowest bidder, we ing widespread. “There are some bright
score higher, but alarm management was can’t standardize on a PLC. We have different spots, however. Several European refineries
second at 15% and predictive maintenance brands, and they have their own protocols, remain committed to adding biodiesel lines,
and security third at 14%,” says Gregory and that becomes a headache. By implement- and these plans have not changed,” accord-
Hale, editor of ISA’s In Tech magazine. Wire- ing OPC, the data all look the same.” ing to ISA’s Cris Whetton.
less’s rank was 13%, and enterprise interop- Construction of stand-alone biodiesel
erability came in at 7% (Fig. 1). Business factors. When asked what they plants is more or less at a standstill, and
Down the road though, the future looks see as the biggest business challenge for the ethanol plants have never attracted the
brighter for wireless. About 22% of those coming year, 45% of survey respondents attention they have in the US, but biodiesel
responding to the survey said that wire- said the recession. The next closest answer integrated with an existing refinery seems
less would be the technology industry users was related to the recession: profitability, to be growing in popularity.
will adopt over the next five years. Asset which came in at 14%. Energy costs and The big growth area is biogas—methane
management was second at 15%, while workforce-development challenges ended produced from biological waste and either
networking and predictive maintenance up at 9%, and the aging out of the work- used locally or injected into a national util-
scored at 14%. Alarm management and force came in at 7%. ity. This is a major growth area in Germany,
security came in at 12%, while enterprise “There is a lot of emphasis on control- Switzerland and Central Europe.
interoperability had 10%. ling costs from what we are told,” accord- Another major growth segment is
Regarding communication, in a turn- ing to Mr. Mitchell. “We will work toward expected to be security systems. In this area,
around from last year, 53% of respondents saving on energy costs. We are focused on wireless solutions are in favor. “For obvi-
said the plant floor is currently able to com- energy cost reduction, and we will do that ous reasons, few are prepared to be specific
municate data through the enterprise to the moving forward.” Between the extra costs about their plans, but as utilities continue
executive suite, while 47% said they did for a plant turnaround that the company to suffer from copper thefts, they are seek-
not. That is the opposite from last year. In has scheduled for this year and the econ- ing wireless solutions, including RFID, for
2008, 47% said they could communicate, omy, it will be tight times at the refinery. access control,” says Mr. Whetton.
while 53% said they did not. “We will not spend
At his refinery, Peter Mitchell, process where we don’t have
controls engineer at the ConocoPhillips to spend,” he says.
Which of these technologies will you adopt over the next five years?
Wireless 22%
Networking 14%
Asset management 15%
Alarm management 12%
Predictive maintenance 14%
Security 12%
Enterprise interoperability 10%
Other 1%
Source: ISA, In Tech, January 2009
FIG. 1 Automation and control professionals respond to a recent FIG. 2 A 400-ton coke drum on barge for delivery to Texas
survey. refinery.

I 17
An ocean of experience.
Unrivalled experience you can trust.

Over the past 30 years, ABB has pioneered the safety system innovations that
have protected people, processes and the environment for generations. Our
installed base of safety systems spans more than 55 countries; ABB protects the
world’s largest oil platform, as well as its most complex pipeline project, and
many other installations. From the very first safety systems in the North Sea
to today's wide variety of leading-edge system options, ABB has developed the
unmatched global expertise along with the solutions and services needed to
effectively make processes safer, more reliable and more efficient.
So why trust your most critical assets to anyone else?
Find out more at www.abb.com/controlsystems.
HPIMPACT
Coke drum delivery Also, 15,000 tons of steel and 180 miles of “We look for just a modest recovery in
piping will be required. real GDP of around 2.5% in the second
marks project milestone half of 2009, as the rebalancing of personal
at Texas refinery Pace of economic decline consumption and savings will take signifi-
TOTAL’s refinery in Port Arthur, Texas, cant time,” says the Conference Board. As
recently achieved a significant project target: forecast to slow a consequence, a 1.7% decline in GDP
the arrival of the centerpieces for its $2.2 in first half of 2009 growth for 2009 as a whole is forecast,
billion Deep Conversion Project. Four mas- The US recession deepened dramatically which is just short of the largest contrac-
sive coke drums—considered to be the heart in the fourth quarter of 2008. However, tion of 1.9% posted in 1982.
of the project—were delivered to the plant according to one recent industry analysis,
from Spain in late January. Each drum is 12 the rate of the economic contraction should Nobel Laureate Chu
stories tall, 32-ft wide and weighs 404 tons. slow in the first half of 2009, and economic
The company invited HP, other media expansion will likely resume in the second selected to head US
representatives and guests to observe this half of the year. The Conference Board, a Department of Energy
construction milestone. nonprofit business and management orga- During his recent Senate confirmation
“This project reflects our strategy of nization, says that its forecast of a 5.9% hearing for Secretary of the US Depart-
investing to enhance the efficiency and annualized decline in real GDP in Q4 2008 ment of Energy (DOE), Steve Chu—an
competitiveness of our large refining hubs reflects across-the-board weakness from the acclaimed physicist and Nobel Laureate—
worldwide, while at the same time reducing negative effect of the escalation in the credit said that boosting development of energy-
our environmental footprint,” according crisis on consumer and business activity in efficient technologies is a critical part of
to Michel Bénézit, TOTAL’s president of the US and abroad. President Obama’s plan to revitalize the
Refining and Marketing worldwide. The worsening labor market, the sharp economy and strengthen energy security.
The Deep Conversion Project includes slide in household net worth, and tighter Dr. Chu, director of Lawrence Berkeley
a 50,000-bpd coker, a desulfurization credit standards resulted in about a 2.5% National Laboratory, pledged to implement
unit, a vacuum distillation units and other decline in real consumer spending, despite the new administration’s goals of increasing
related components. very steep and early holiday discounting and research and development of new energy
The new units will increase the facility’s a rapid decline in the consumer price index. technologies, developing fuel-efficient vehi-
deep-conversion capacity and expand its External demand for US exports also cles and increasing the energy efficiency of
ability to process heavy and sour crude oil. dropped precipitously as the financial crisis buildings and appliances.
With the upgrades, 3 million tons/yr of spread globally and the economic recession “We are very fortunate to have a nomi-
ultra-low-sulfur automotive diesel will be deepened among major trading partners. nee of Dr. Chu’s high caliber to take on
added to the refinery’s production, raising Companies greatly reduced their inventory these responsibilities. He will bring to the
total output of all products combined to levels in Q4 by about $67 billion. “Invento- job the keen scientific mind of a physicist
about 12 million tons/yr. Project commis- ries will continue to be a drag on growth in and Nobel Laureate,” said US Senator Jeff
sioning is scheduled for 2011. the first half of 2009, but since more of the Bingaman (D-NM), speaking at Dr. Chu’s
The undertaking is using the latest inventory correction occurred in Q4 than confirmation hearing.
generation of coker technology. TOTAL is we previously forecast, they will help limit Dr. Chu was a committee member of The
adapting refining operation to meet pres- the contraction of growth in Q1 and Q2,” American Physical Society that produced the
ent and future transportation fuels mar- according to the Conference Board. report, Rising Above the Gathering Storm.
ket. “The refiner must evolve to remain “The over-arching message of that
competitive,” Mr. Bénézit said. This proj- Slowing slide? Despite the consider- report is simple: The key to America’s
ect increases the refinery’s complexity and, able downside risks that exist, the fourth- prosperity in the 21st century lies in our
according to Mr. Bénézit, project payback quarter 2008 could mark the deepest part ability to nurture and grow our nation’s
should be achieved in one year. of the recession. This analysis suggests “a intellectual capital, particularly in science
good likelihood” that the US economy will and technology. As the largest supporter of
New units. The core project involves post a modest recovery by the second half the physical sciences in the US, the Depart-
constructing the following new units: of 2009. Financial market conditions are ment of Energy plays an essential role in the
• Coker (deep conversion unit) showing some signs of improvement, led training, development and employment of
• Vacuum distillation unit to prepare by a noticeable recovery in the short-term our current and future corps of scientists
the coker feed money markets and a narrowing in invest- and engineers.”
• Distillate hydrotreater ment and noninvestment grade corporate In 1997 while at Stanford University,
• Coker naphtha hydrotreater bond yields. Dr. Chu was one of three scientists to win
• Hydrogen purification–PSA Significant monetary and fiscal policy the Nobel Prize in physics for developing
• Sulfur recovery. easing is providing much-needed capital methods of cooling and trapping atoms
In addition, the power supply of the and bolstering confidence, though a high with lasers—work that he carried out at the
refinery will be modernized by connecting degree of risk aversion keeps financial con- former AT&T Bell Laboratories.
the new entity to the 230-kV network. The ditions far from normal. At the same time, Dr. Chu is the first Nobel Laureate to be
upgrade will use about 70,000 cubic yards concerns about a rising deficit and govern- confirmed as a Cabinet member. He suc-
of concrete—more than the quantity used ment debt are mounting and will likely ceeds Samuel W. Bodman, who held the
to construct the Empire State Building. damper future economic growth. post since January 2005. HP
I 19
leave your mark
on tomorrow’s
energy solutions
ExxonMobil is seeking experienced engineers with proven leadership skills for refining and
chemical positions in Beaumont, Texas. Qualified individuals will have a B.S. or higher in
Chemical, Mechanical, or Electrical Engineering; relevant experience; a demonstrable history
of effective leadership in a team environment; and extensive expertise in specific areas:
• Delayed Coker Process Engineer - B.S. Chemical or

Mechanical Engineering (job # 7481)
• Continuous Catalytic Reformer Process Engineer
- B.S. Chemical Engineering (job # 7481)
• Light Ends Process Engineer (fractionation, alkylation,
isomerization) - B.S. Chemical Engineering (job # 7481)
• Refinery Utilities Engineer (gas turbine generators, boilers,
water treating) - B.S. Chemical or Mechanical Engineering
(job # 7481)
• Energy Conservation Engineer (combustion, heat exchanger,
and steam system management, energy projects) - B.S.
Chemical, Mechanical, or Electrical Engineering (job # 7481)
• Refinery Process Control Engineer - B.S. / M.S. Chemical
Engineering (job # 7218)
• Olefins/Aromatics Process Control Engineer - B.S. Chemical
Engineering (job # 7218)
• High Pressure Machinery Engineer for polyethylene plant
- B.S. / M.S. Mechanical Engineering (job # 7479)
• Instrument Engineer (general unit support, compressor
specialist, PLC coordinator, or large project support)
- B.S. / M.S. Electrical Engineering (job # 7220)
• Fixed Equipment Engineer for polyethylene plant - B.S. /
M.S. Mechanical Engineering (job # 7423)
Please apply online at exxonmobil.com/ex to the job numbers listed above.

(Note: please apply to the two jobs that most closely match your skills and
interests, as you are limited on the number of jobs to which you may apply.)
Additional information on position duties is available online.
Exxon Mobil Corporation An Equal Opportunity Employer
TM
Taking on the world’s toughest energy challenges.
HPINNOVATIONS
SELECTED BY HYDROCARBON PROCESSING EDITORS
[email protected]
Regenerable SO2 scrubbing thus eliminated, and effluents are reduced Fig. 1 illustrates how the regenerable
eases environmental pressures to a minimum. Furthermore, the high amine scrubber can be integrated into an
To manage growing strategic pres- capacity and selectivity of the absorbent existing three-stage SRU that is designed
sures from green fuels and environmental reduce capital costs. for 97% conversion efficiency at the end
issues, refiners will be required to direct This patented technology uses an aque- of catalyst run conditions. In this case,
more attention to their refinery total sulfur ous amine solution to achieve high-effi- operating costs do not include natural gas
balance. Non-regenerable sulfur dioxide ciency selective absorption of SO2 from consumption and steam production in the
(SO2) scrubbing systems will increase costs a variety of gas streams. The scrubbing tail-gas thermal oxidizer.
as expenses for reagents such as sodium byproduct is pure water—saturated SO2 Extensive flue gas cooling is required
hydroxide, lime or limestone increase. gas is recovered by steam stripping, which to chill the gas to absorber conditions and
Further, tighter environmental controls is low-quality heat. remove water formed by the Claus reaction.
will likely limit disposal of gypsum to land- The scrubbing systems have been oper- The prescrubbing system must purge 44
fill or to disposal of sodium sulfate into ating in various refining units, including: gpm, or 7.3 tons of water per ton of SO2
refinery wastewater streams. Regenerable • Fluid catalytic cracking unit and fluid captured by the tail-gas system. On an SRU
SO2 scrubbing systems can help ease many coker carbon monoxide boiler SO2 scrubber basis, this translates to 0.4 tons of water per
of the environmental and market-induced • Claus sulfur recovery unit (SRU) ton of sulfur directed to the pit.
pressures that are associated with greater • Capture-SO2 from flue gas generated Select 1 at www.HydrocarbonProcessing.com/RS
use of high-sulfur crude oils. by resid-fired crude unit process heaters
The CANSOLV SO2 Scrubbing Sys- and utility boiler systems. Gas analyzer sets new standard
tem, operating commercially since 2002, is in oxygen measurement
claimed to be a proven regenerable amine SRU tail-gas scrubbing. To manage Servomex has introduced the SERVO-
technology that removes SO2 from various higher sulfur loadings and process lower TOUGH Oxy oxygen gas analyzer. It is
gas streams found in refineries and petro- sulfur-content transportation fuels, revamp- claimed to offer an exceptional range of
chemical facilities. The system is regenera- ing the refinery will require adding an SRU industry-standard options and three
ble—meaning that the chemical absorbent tail-gas cleanup system. This can also be unique, groundbreaking functions. The
is not consumed within the process. The satisfied by installing the CANSOLV SO2 analyzer is expected to set new flexibility,
high costs of consumable absorbents are scrubber as part of the SRU expansion. stability and reliability standards from a
single, cost-effective unit.
To stack As well as fault and calibration histo-
ries, all units offer NAMUR-compliant
CANSOLV battery relay functions, allowing two concentra-
Amine purification
limits
unit
tion alarm levels and maintenance-required
service in progress, and instrument fault
messages to be communicated remotely.
Amine
absorber A comprehensive Modbus protocol allows
SO3 removal remote communication and unit interroga-
Regenerator
tion as standard via RS485, as well as an
option for Ethernet connectivity.
Quench/cooling Makeup
water Auto-validation and auto-calibration
functions allow users complete flexibility
for unmanned or remote operation, or to
Purge water to generate maintenance and reliability sched-
water treatment ules using trending information. Stainless-
Steam Steam Steam Steam steel pipe work, automatic range change,
Fuel Steam
Acid gas As HP editors, we hear about new products,
H2S, SO 2 patents, software, processes, services, etc.,
Thermal that are true industry innovations—a cut
Reaction
oxidizer furnace above the typical product offerings. This sec-
Air Air tion enables us to highlight these significant
developments. For more information from
Sulfur Sulfur Sulfur Sulfur
these companies, please go to our Website
FIG. 1 SRU tail-gas cleanup unit can be integrated into an existing three-stage SRU. at www.HydrocarbonProcessing.com/rs and
select the reader service number.

I 21
HPINNOVATIONS
PV Elite
VESSEL &
fixed background gas compensation and rials and causes environmentally harmful EXCHANGER
measurement filtration are also standard. sulfur dioxide emissions.
The Oxy introduces three unique EN 14214 and ASTM D6751 stan- ANALYSIS
options: dards have been introduced specifying the
• An innovative, fully heated sample requirements for biodiesel and its analysis.
compartment removes the requirement for These documents require that the con-
a sampling conditioning system on all sam- centrations of elemental contaminants in
ples with a dew point up to 50°C. Respon- biodiesel be regularly monitored and spec-
sible for up to 80% of failures in compa- ify the method for its analysis. The aim is to
rable units, sample conditioning failure is ensure optimum engine performance and
a major cause of unplanned downtime. The reduce environmental impact.
heated sample compartment design reduces Traditionally, axial-view ICPs have been
this risk of downtime by removing coolers, the configuration choice for ICP emission
dryers and other conditioning devices. spectrometers used to perform biodiesel
• A unique flow sensor has been placed
after the measurement outlet, guarantee-
ing accurate flow alarm settings for all uses
analyses due to lower detection limits.
Owing to the robust nature of its dedicated
radial view plasma and the elimination of
Easy
including safety applications.
• A novel integrated pressure compen-
sation system not only compensates for
barometric pressure but also for back pres-
carbon-based emission interferences associ-
ated with the axial view configuration, the
new spectrometer’s radial view is claimed
to be a powerful alternative, consider-
Accurate
sure variations from flare stacks, enabling
emission compliance targets to be easily
met. Both the flow sensor and pressure
ably increasing analytical sensitivity for
important elements such as phosphorus
and sulfur.
Reliable
compensation system technologies report This configuration demonstrates improved
What makes PVElite one of the
via the instrument’s standard communica- detection limits for lower concentrations of
fastest growing vessel & exchanger
tion options, providing all measurement and samples, being capable of providing accurate,
safety benefits without the need to install dependable phosphorus, sulfur and potassium analysis solutions on the market?
additional devices and cost-hungry cabling. analysis. This is a crucial benefit as, according Because of design and analysis features
such as:
Potential applications for the analyzer to regulations, detection limits must be 10
include usage in process control, safety times below the regulated concentration levels • Intuitive user interface
• ASME VIII 1&2, EN 13445, PD5500 codes
critical oxidation such as ethylene and to provide sufficient margin for ensuring a
• Analysis to TEMA standard
propylene oxide, flare stack analysis, prod- sensitive measurement. • International wind/seismic codes
uct purity, feedstock cleanup and inerting • Stack design
or blanketing. Other features. Additionally, the iCAP • Fitness for Service capabilities
Select 2 at www.HydrocarbonProcessing.com/RS 6000 spectrometers have full wavelength • Component calculations with CodeCalc®
coverage from 166 nm to 847 nm with built-in
• Ability to mix and match units for analysis
Biodiesel analysis uses full frame capability, offering full spectrum
and reporting
radial plasma view trend analysis and contamination identifica- • Comprehensive output & reports
Thermo Fisher Scientific Inc. has incor- tion between batches of biodiesel produced. • Bi-directional links to CADWorx®
porated unique capabilities in the iCAP Their advanced optical design enables Equipment Module
6000 Series of ICP emission spectrome- improved resolution and detection limits.
ters to achieve dependable monitoring of The systems are fitted with a fourth-gen- PVElite delivers!
elemental contaminants in biodiesel. The eration CID detector. This provides a wide Contact us to find out how you can improve
dedicated radial plasma view system config- dynamic range, resistance to saturation and your design engineering efficiency.
uration is claimed to provide enhanced ana- greater detection capability. The new series
lytical capabilities for important elements incorporates fully automated wavelength
such as sulfur and phosphorus. Addition- calibration and offset correction capabilities
ally, the enhanced matrix tolerance torch for excellent long-term stability.
and swing frequency RT generator easily The instrument’s distributed purge sys-
handle organic matrix samples and ensure tem offers reduced gas consumption and
improved stability. improved performance for elements such
Most biodiesel production plants use as sulfur and phosphorus that emit light in
Plant Focused.
plant oils as a starting material for produc- the ultraviolet spectrum region. The spec- Industry Driven. SM
©2007 COADE, Inc.
tion. However, these plants usually have trometer’s ergonomic design—with a large,
relatively high phosphorous content. This wide-opening door—enables easy access +1 281-890-4566 • [email protected]
is undesirable in fuels as it can lead to cor- to the sample compartment and peristaltic
rosion of mechanical engine components. pump. This makes routine maintenance www.coade.com
Sulfur also affects engine wear if present in easier and faster. DOWNLOAD FREE DEMO
excess concentrations in the starting mate- Select 3 at www.HydrocarbonProcessing.com/RS

HPIN CONSTRUCTION
BILLY THINNES, NEWS EDITOR
[email protected]
North America refinery. CB&I’s scope of work for the proj- France. With capacity expanded by
Total’s refinery in Port Arthur, Texas, ect includes the engineering, procurement 210,000 metric tpy, the 600,000-met-
recently added some equipment as part and fabrication of the hydrotreating unit, ric tpy unit will be one of the largest in
of an ongoing $2.2 billion upgrade. The which removes sulfur from diesel by utilizing Europe. The unit’s startup is part of the
upgrade, known as the Deep Conversion a catalyst in the presence of hydrogen. industrial restructuration project launched
Project, includes a 50,000-bpd coker, a by Total Petrochemicals in France in the
desulfurization unit, a vacuum distillation South America spring of 2007. Central to this plan, Total
unit and other related units. The new units INEOS Technologies has granted two Petrochemicals’ styrene business in Europe
will add 3 million tpy of ultra-low-sulfur polyethylene technology licenses to Poli- has been rescaled and consolidated at the
diesel to the refinery’s current production. mérica S.A. These plants will form part of Gonfreville complex, resulting in the shut-
The project should be complete in 2011. Polimérica’s cracker and derivatives complex down of the Carling unit in France. This
in José, Venezuela. Startup of the complex reduced overall styrene production capacity
Enerkem Inc.’s plant in Westbury, is planned for 2013. by 120,000 metric tpy.
Quebec, Canada, recently entered a startup The first of the two new facilities will Project capital expenditure amounted
phase with the production of its clean-con- be a 430,000-mty gas phase polyethyl- to €320 million, including €20 million to
ditioned syngas. Construction on the plant ene plant using INEOS swing gas phase adapt the site infrastructure and improve
began in October 2007 and the facility was technology to produce linear low density safety and environmental standards. Due
mechanically complete in December 2008. polyethylene (LLDPE) and high density in part to its new reactors, the styrene unit’s
Once the facility begins production, it will polyethylene (HDPE). The other will be energy efficiency has increased 30%, thereby
produce liquid fuels and green chemicals a 400,000-mty slurry polyethylene plant reducing carbon emissions from styrene pro-
using renewable, non-food, negative-cost using INEOS slurry technology for the duction processes by a similar percentage.
feedstock, like wood from used electricity production of HDPE.
poles. Production is forecast for 1.3 million Burckhardt Compression has an order
gpy of second-generation ethanol. Europe from a refinery in northern Italy to deliver
Total Petrochemicals recently started two process gas compressors for a new mild
Praxair, Inc. has a hydrogen supply con- up a revamped styrene unit at its petro- hydrocracker unit. The contract comprises
tract from Dynamic Fuels, LLC. Dynamic chemicals facility in Gonfreville-l’Orcher, two multiservice makeup process gas com-
Fuels will use hydrogen supplied by Praxair to pressors that are equipped with a monitor-
produce renewable fuels from non-food-grade TREND ANALYSIS FORECASTING ing and diagnostic system and the recycle
animal fats produced or procured by Tyson Hydrocarbon Processing maintains an arrangement in three compression stages.
extensive database of historical HPI proj-
Foods. Diesel and jet fuels will be produced at ect information. Current project activity
The compressors will be used for the pro-
Dynamic Fuels’ Geismar, Louisiana, produc- is published three times a year in the HPI duction of ultra-low-sulfur diesel and are
tion facility by using fats such as beef tallow, Construction Boxscore. When a project driven by 2,500-kW electric motors. They
pork lard, chicken fat and used greases. is completed, it is removed from current are scheduled to be delivered in July 2009.
listings and retained in a database. The
Dynamic Fuels’ $138 million plant is database is a 35-year compilation of proj-
The plant will start the production of clean
currently scheduled to begin production ects by type, operating company, licen- diesel at the beginning of 2010.
in 2010, with a total capacity of 75 million sor, engineering/constructor, location, etc.
gallons of fuel per year. Many companies use the historical data for Foster Wheeler Ltd.’s Global Power
trending or sales forecasting.
Group has been awarded a contract for a
The historical information is available in
Jacobs Engineering Group Inc. has a comma-delimited or Excel® and can be cus-
heat recovery steam generator (HRSG) by
contract from a major oil and gas company tom sorted to suit your needs. The cost of UTE IBERESE-SOMAGUE. The boiler
for the engineering of a thermal facility the sort depends on the size and complex- will be integrated in a cogeneration plant
in northeast Alberta, Canada. Jacobs will ity of the sort you request and whether a that Repsol is constructing at the Sines
customized program must be written. You
perform pre-engineering design specifica- can focus on a narrow request such as the
refinery in Portugal. Foster Wheeler will
tions (pre-EDS), EDS, detailed engineering history of a particular type of project or design, supply and erect the HRSG, and
and procurement services for the in-situ oil you can obtain the entire 35-year Boxscore will also provide startup supervision for
sands central processing facility. Engineer- database, or portions thereof. the HRSG, which will be coupled to a Sie-
ing activities began November 2008 and Simply send a clear description of the data mens SGT-800 combustion turbine, with
Jacobs is scheduled to complete its scope you need and you will receive a prompt a total installed ISO rating of 47 MWe
cost quotation. Contact:
in February 2011. (gross megawatt electric). The HRSG will
Lee Nichols
P. O. Box 2608
produce high and low-pressure steam for
CB&I has a contract, valued in excess of Houston, Texas, 77252-2608 the refinery process. Commercial operation
$50 million, to design and fabricate a distil- Fax: 713-525-4626 of the HRSG is scheduled for the second
late hydrotreating unit for a North American e-mail: [email protected]. quarter of 2010.
I 25
HPIN CONSTRUCTION
Middle East at Qatar Gas 1, as well as extend the life expected to begin startup in 2013.
Invensys Process Systems (IPS) has cycle of the overall control system there. WorleyParsons’ scope of work includes
signed a multimillion dollar contract with front-end engineering design (FEED) and
Qatargas to complete a major automation Saudi Aramco Mobil Refinery Co., Ltd. full responsibility for engineering, procure-
upgrade at the Qatar Gas 1 facility in Ras (SAMREF) has selected WorleyParsons to ment and construction of the facilities.
Laffan Industrial City, Qatar. Under the execute its Clean Fuels Project at Yanbu Al- Dependent on finalization of the scope of
terms of the contract, IPS will upgrade con- Sinaiyah, Saudi Arabia. The project encom- work details, WorleyParsons’ services contract
trol processors, gateways, local area networks passes significant modifications to SAMREF’s value could be as high as $ 400 million.
and network security. The upgrade will give refinery to comply with future mandatory
Qatargas improved compatibility between sulfur levels of 10 parts per million in gasoline Technip has an EPC contract with
different generations of system components and diesel. The phased construction project is Middle East Oil Refinery (MIDOR), esti-
mated at approximately €43 million, for
the expansion of the delayed coking unit of
its refinery in Alexandria, Egypt. Engineer-
MORE THAN JUST Agriculture ing, procurement and supply of equipment
SHARING YOUR VISION and materials will be delivered on a lump-
Agri-food sum basis; construction activities will be
TOGETHER, charged on a reimbursable basis.
Chemicals
WE CAN COMPLETE IT. and Petroleum
The delayed coking unit, based on
ConocoPhillips technology, will have a
Environment production capacity of 30,000 bpd. It is
scheduled to be delivered by the third quar-
Facilities ter of 2010.
and Operations
Maintenance Asia-Pacific
Industrial LyondellBasell Industries recently
and Manufacturing started up its new polypropylene (PP) com-
pounding facility in Nansha, China, with a
Infrastructure nominal capacity of 15,000 tpy. The new
Mining and facility is operated by Guangzhou Basell
Metallurgy Advanced Polyolefins Co., and supplies
polypropylene composites and alloy materials
Pharmaceuticals to the automotive and appliance industries.
Power
Black & Veatch’s LNG process contrib-
Telecommunications uted to the development of the LNG facility
in Erdos, China, which attained full pro-
duction capacity of 200,000 metric tpy in
December 2008. A second facility that Black
& Veatch worked on, in Zhuhai City, China,
has also begun commercial operation.
Four additional plants are planned in
SNC-Lavalin designs, develops and delivers leading engineering, central Sichuan Province, central Shaanxi
construction, infrastructure and ownership solutions worldwide. We Province, northwestern Gansu Province
listen carefully to you, and the communities you serve, while striving and the northwestern Xinjiang Autono-
mous Region. Upon completion, the six
for excellence in our commitment to health, safety and the environ-
facilities will supply a total of approximately
ment. We have the global versatility and technical expertise to meet 1.2 million tpy of LNG.
your expectations and complete your vision. www.sncl.us
Africa
Shell Global Solutions International BV
has a contract with Oilmoz Lda to design a
refinery. Oilmoz Lda plans to build an $8 bil-
lion, 350,000 bpd oil refinery in the Maputo
province of Mozambique. Shell Global will
start off as technical adviser for the project
SNC-Lavalin Engineers & Constructors Inc. and will later become technical partner. Once
9009 West Loop South, Suite 800 • Houston, Texas 77096 • USA • 713-667-9162 • [email protected]
the refinery is completed, it will be the first in
North America Latin America Europe Africa Eurasia Asia Middle East Oceania Mozambique in over 24 years. The comple-
tion date is scheduled for 2014. HP
26
Select 65
Select 401atatwww.HydrocarbonProcessing.com/RS
Improve Plant
Profitability and
Maximise Your Site’s
Potential with KBC…
OpX – Energy Efficiency Initiative, European Refiner.

In the first phase, KBC conducted a review of the Mesoamérican hydrocarbon markets,
KBC holds a long-term working relationship with a client, which operates a
medium-complexity refinery along the coast of Romania. In an effort to improve the
performance and profitability of its operations, the refiner decided to undertake an
Energy Efficiency Improvement Initiative. The work consisted of:
U Benchmarking of Refinery Energy Performance
U Gap Analysis of Areas of Inefficiency
U Fired Heater Assessment
U Steam/Power System Modelling and Optimisation
U Selective Process Unit Energy Optimisation, including:
- Selected Heat Integration (Pinch) Studies
- Process Unit Simulation (using KBC Petro-SIM™) and Optimisation
U Equipment-level Analysis (furnaces, turbines, exchangers, fouling)
After the analysis was complete, KBC presented recommendations to the client, and within
four months, the refiner reported that the benefits implemented and achieved amounted to
USD$4.3 Million/year. KBC was also able to help the client move from the 3rd quartile to For more information on how KBC can help you
the 2nd quartile of energy efficiency among over 200 other refiners surveyed by KBC
around the world. achieve Operational Excellence, contact us at
KBC has performed successful Operational Excellence (OpX) programs for clients around AMERICAS +1 281 293 8200
the world in the areas of:
EMEA +44 (0)1932 242424
U Operational Planning U Reliability, Availability, & Maintenance
U Process Optimisation U Human Performance Improvement ASIA +65 6735 5488
U Energy U Software Solutions
U HSE [email protected] U www.kbcat.com
INSTRUMENTS AND NETWORKS SPECIALREPORT
Wireless networks improve

refinery operation
Smart instruments and secure wireless communications
enable enhanced operations and asset management
G. MARTIN, Emerson Process Management, Austin, Texas
M
odern wireless technology • The same company installed a
■ The low installed cost, reliability,
provides valuable auto- 45-transmitter wireless monitoring
mation options for oil security and ease-of-use of the network in a tank farm at a tech-
refineries today—improving work- nology center, avoiding the cost of
force productivity, safety and plant newest wireless networks are engineering and constructing a wired
security. Wireless communications system to obtain a continuous stream
enable access to all assets in the refin- causing increased awareness of their of data on suction and discharge pres-
ery, including instruments, valves, sures, levels, flow and temperatures.
controllers, equipment, cameras possibilities in refineries, bringing • At Hunt Refining, Tuscaloosa,
for safety and security, and people. about some innovative applications AL, three wireless temperature trans-
Wireless mesh access points use open mitters on a single hot asphalt tank
standards for compatibility and 128- with excellent results. help identify “hot spots” that can
bit encryption for security. Wireless lead to roof corrosion which could
field devices network themselves using a self-organizing mesh that cause a roof failure costing as much as $200,000.
automatically reroutes signals around obstructions. • At the same location, a wireless device monitors the tempera-
ture of cooling water being returned to the local river to assure
Why wireless? A wireless solution eliminates “blind spots” compliance with environmental regulations.
in the plant—operation areas that have been either technically • Vibration readings on five critical pumps in a hazardous area
or economically unreachable with wires. While these areas are at a Midwestern refinery are transmitted wirelessly and integrated
often not critical, they do play a major role
in overall refinery performance and safety.
Smart wireless networks also provide
convenient access to diagnostics that already
exist in hundreds of plant devices that have
no way to deliver them for operations use.
Wireless can communicate the information
to operators through Web-based portals.
And the clipboard walk-arounds, conducted
by plant staff because there was simply no
other way to get the data back to operations,
are replaced by automated solutions.
The low installed cost, reliability, secu-
rity and ease-of-use of the newest wireless
networks are causing increased awareness
of their possibilities in refineries, bringing
about some innovative applications with
excellent results. For example:
• The 225,000 bpd BP Refinery at
Cherry Point, WA, installed the first indus-
trial wireless mesh field network in 2006,
which continues to operate reliably, elimi-
nating time-consuming operator rounds in FIG. 1 Wireless in a digital plant.
the field.
I 29
SPECIALREPORT INSTRUMENTS AND NETWORKS
key components of TSMP that contribute

to end-to-end network reliability are:
• Time-synchronized communication
• Frequency hopping
• Automatic node joining and network
formation
• Fully-redundant mesh routing
• Secure message transfer.
TSMP networks are robust/tolerant to
almost all interference and coexist with
other wireless networks. Robust security
is designed in. Demonstrated reliability is
greater than 99%.
Self-organizing wireless field networks
FIG. 2 Self-organizing field network.
can be easily installed and deliver significant
value without the need for investing in a
plantwide wireless infrastructure.
with the DCS and plant historian. A serious bearing issue was
identified within 24 hours of startup, avoiding a process upset. Wireless network security. At the wireless field network
level, robust security is provided through advanced, standards-
Wireless digital architecture. The wireless digital plant based encryption as well as authentication, verification, key man-
(Fig. 1) functions like a conventional wired plant, but with one agement and antijamming techniques.
major difference. Communications between devices are trans- Smart wireless solutions employ end-to-end 128-bit encryp-
mitted as radio frequency signals. This makes the wireless digital tion using the advanced encryption standard (NIST standard
plant very cost-effective because complex wiring racks including FIPS-197). For authentication purposes, each gateway maintains
fiber-optic runs are not necessary. a “white list” of devices allowed to communicate with it, and indi-
And the wireless technology is scalable. Field or plant network vidual devices accept messages only from a previously identified
applications are easily added in existing facilities, or to brownfield gateway or from other gateway-validated devices.
or greenfield projects. These solutions for process and plant man- Separate “join” and “network” keys can be set to automati-
agement applications install easily and operate reliably, improving cally rotate or be changed on demand. Implementing the Wire-
productivity, safety and operational efficiency. lessHART standard adds “session” keys for communication
between two network devices so that other devices can’t “listen
Open-standard solutions. Smart wireless solutions are in”. These can be rotated as well.
based on open standards—WirelessHART for the field network Message integrity codes (MICs) are used to verify messages,
and IEEE 802.11 a/b/g Wi-Fi for the plant network. both per-hop and end-to-end. Antijamming techniques such as
The 802.15.4-based WirelessHART standard calls for self- direct sequence spread spectrum (DSSS) with channel hopping plus
organizing technology that delivers high communications reli- multipath routing help sidestep noise sources, whether malicious or
ability in wireless field networks. not. And gateway-to-host security leverages well-known standards
The Class 1 Division 2 wireless access points used in smart such as SSL as well as complete encryption/authentication.
wireless plant-level networks are compliant with IEEE 802.11i and At the wireless plant network level, security is fundamen-
Wi-Fi Protected Access 2 (WPA2), which employs hardware-based tal to the unified wireless network. The standards-based self-
Advanced Encryption Standard for wireless communications. defending network solution provides confidence that the plant
and business data will remain private and secure. Threat-control
Self-organizing networks. At the heart of the smart wireless capabilities control and contain known and unknown threats,
system is the self-organizing mesh network. Secure and infinitely and network admission control helps to enforce organizational
configurable, the self-organizing network ensures an adaptive,
flexible approach to wireless that defies the “canyons of metal”
that define most plants.
Unlike many approaches to in-plant wireless that require direct
line-of-sight between the instrument and the communications
gateway, the smart wireless approach ensures the greatest network
integrity by allowing devices to communicate with each other.
This means there is no single point of failure; every device serves
as a network connector. In the event a temporary obstruction
blocks a direct connection, the network automatically reroutes
the signal to an adjacent device, ensuring network reliability and
data integrity.
Self-organizing mesh networks (Fig. 2) have demonstrated
high reliability in the field. They use the IEEE 802.15.4 time-syn-
chronized mesh protocol (TSMP) with added channel hopping. FIG. 3 Several wireless devices are available.
TSMP can be supported by both 900 MHz and 2.4 GHz. Five
30
I MARCH 2009 HYDROCARBON PROCESSING
Innovative wireless applications grow at BP

BP is finding more and more ways been able to in the past. We really hope flow and temperatures. New wireless func-
to use wireless field devices at its Cherry our wireless technology will be a principal tions are installed as they become available
Point refinery in Washington, through- tool in maintaining plant availability while for refinery-wide applications. The real-
out the tank farm in its R&D facility in expanding our flexibility to meet fuel specs world environment in a pilot-scale opera-
Naperville, Illinois, and at other refineries and ever-changing array of feedstock.” tion provides hands-on experience for the
around the world. Wireless has found a natural home at process engineers and valuable feedback
At the 225,000 bpd Cherry Point refin- BP’s Naperville R&D facility, a world-class for refinery management. Options for
ery, a 15-transmitter wireless installation technology center including a model tank refinery process optimization and shar-
in the calciner unit monitors bearing and farm feeding pilot plants that develop pro- ing of wireless automation technology are
calciner coke temperatures to help prevent cessing options for BP refining worldwide. thereby shared globally by the Refining
fan and conveyor failure. Fans can cost up Following the success of wireless at Cherry Technology team.
to $100,000 to repair but, more impor- Point, BP installed a 45-transmitter wire- “Wireless is an important enabler for
tantly, can be down for up to 10 days with less network to monitor the Naperville ‘refinery of the future’ technologies,”
associated production losses. This wireless tank farm. Operational for more than one commented Mark Howard, commercial
network, believed to be the world’s first year, this installation has provided strong technology manager for BP. “It helps us
industrial application of the self-organizing operational experience and a platform for deploy the sort of instrumentation, sen-
wireless mesh technology in 2006, contin- testing the technology, leading to signifi- sors and analytical devices that we need
ues to operate reliably while eliminating cant use of wireless at other BP refineries for condition monitoring to support pre-
operator rounds in the field. throughout the world. dictive maintenance, tracking feedstock
Cherry Point has since expanded the “The wireless devices allow our opera- through the value chain and a host of other
use of wireless to 35 transmitters includ- tors to be more efficient, collecting data applications. Wireless is a very important
ing tank farm and utility applications, and from one central point as opposed to vehicle for getting instrumentation into
a smart wireless gateway in the diesel unit walking around the tank farm and record- places where wired devices would be too
has made it ready for wireless motes. ing all the values,” comments a BP rep- expensive or frankly not very practical.
“The principal advantage we see around resentative. “The other advantage of the “Looking ahead, we support the move
wireless is the ability to accumulate and wireless devices is that they supply data toward standards such as WirelessHART,”
analyze a much greater array of data than continuously for recording in our histo- Howard said. “We like being able to access
would otherwise be economically possible,” rian, allowing us to see what is happening new wireless transmitters as quickly as we
said Michael Ingraham, technical manager in the tank farm at any time of the day.” can deploy them, and we’re getting very
for the Cherry Point refinery. “Wireless The Naperville tank farm network uses good, robust operation. We look forward
enables us to get more data more efficiently, wireless transmitters to monitor pump to a greater range of instrumentation
more economically than we ever have suction and discharge pressures, levels, becoming available.”
security policies to allow only trusted end-point devices to access

the network.
The wireless access points used in smart wireless plant-level
networks are compliant with IEEE 802.11i and Wi-Fi protected
access 2 (WPA2) that employ hardware-based advanced encryp-
tion standard encryption for wireless communication.
Wireless field networks. A variety of wireless field devices

exist, including: pressure, flow, level, temperature, vibration, cor-
rosion, valve position monitor, multi-input temperature, discrete
switch, HART upgrade module and wireless device router. Some
of these are illustrated in Fig. 3.
Wireless instruments can be widely and remotely distributed
throughout a plant, across roads and ponds, or on mobile platforms
like railcars, barges, or trucks where traditional wired data collection
is not feasible. Only small amounts of bandwidth are used for high- FIG. 4 Wireless plant network.
priority bursts of data from each device serving as a transmitter. The
self-organizing mesh network continuously monitors signals for signs
of degradation and repairs itself as necessary, automatically finding the Wireless plant network. There’s a lot more to a plant than
optimum communication route to the network gateway. If a tempo- what goes on inside the pipes and process vessels. Just as impor-
rary obstruction blocks a connection, signals are rerouted via adjacent tant as the products sold are the people who make them—and
wireless devices, which act as transceivers. In this way, connectivity is the information they use to do their jobs. Not just in offices or
maintained while achieving high data transmission reliability. control rooms, but out in the field. Applications include video,
I 31
voice, mobility and tracking. The wireless plant network is illus-

trated in Fig. 4.
Wireless technology makes it easier to put those people in
touch with the information they need, wherever they are.
• Workers can access desktop applications and perform tasks
wherever they are—including viewing and responding to alarms
from the field.
• The locations of personnel and physical assets in the plant
are tracked at all times—especially useful for safety mustering.
• Messages can be broadcast to specific groups of workers
wherever they are.
• Security systems track and ensure authorized plant access.
• Video systems not only patrol the fence line, but keep a cost- FIG. 5 Grow the wireless network as needed.
effective eye on the process.
At the plant-network level, two types of wireless applications
offer significant benefits: workforce productivity and business and helping security managers identify potential vulnerabilities and
plant management. improve systems.
Wireless applications also enable one to remotely monitor
Workforce productivity. Wireless technology empowers plant security breaches or onsets of hazardous situations. Wireless
plant workers by giving them instant access to information wher- monitoring of these situations gives plant security and operations
ever they are. teams time to prepare and dispatch personnel that are suitably
Operators can perform many of their duties from the comfort equipped for the situation.
and safety of the control room—but there are still times when And wireless location technologies allow quick access and track-
they have to go out into the field to collect data, check on equip- ing of inventory and valuable assets—even workers—moving
ment or just see firsthand how the plant is running. inside and outside the plant. Time spent looking for assets can
Put a ruggedized wireless PC in their hands, and now they can be dramatically reduced, which can have significant benefits dur-
remotely access control and asset- ing major turnarounds, emergen-
management systems to immedi- ■ Today’s wireless technology enables cies and new construction projects.
ately relate what they see to what’s Being able to quickly locate each
happening in the process—and a top-down or bottom-up approach. worker also offers safety and pro-
respond as needed. That includes ductivity benefits.
viewing and acknowledging alarms Process manufacturers can begin at
no matter where the operator is. Start anywhere and grow
Communications improve, too.
the plant level and work down or at the wireless system. Today’s
While many plant workers already the field and work up, wherever the wireless technology enables a top-
use an older wireless technology— down or bottom-up approach. Pro-
walkie-talkies—for short-range highest priority needs are. cess manufacturers can begin at the
communications in the field, com- plant level and work down or at the
bining a plantwide wireless broadband network with Voice over field and work up, wherever the highest priority needs are. While
Internet Protocol (VoIP) technology can extend communication the reasoning and justification varies widely, a field-level network
reach while also enabling “smart” communications. For example, of measurement devices is often chosen as the first foray into the
one can broadcast messages to specific teams based on the IP wireless world. As a result of the flexibility, scalability and ease of
address of each worker’s radio. use, the process industries already benefit from proven experience
Maintenance workers can also benefit. Wireless tools such over several years and hundreds of wireless installations. This suc-
as handheld communicators let them access maintenance work cess supports starting now and growing.
orders, instructions and other information on the spot, and to Fig. 5 illustrates wireless flexibility and scalability. The small
immediately track or report inspections, tests and repairs. image on the left shows a tank level application, a common start-
ing point for a wireless system because of remote locations and
Business and plant management. Wireless applications associated costs of implementing a wired solution. Subsequent
such as personnel and asset tracking, as well as wireless video surveil- applications can be added as shown, all being automatically linked
lance for security and safety, have changed the way offices, hospitals, via the self-organizing network. HP
warehouses and retail stores operate. Now they can deliver business
and plant-management benefits in process operations, too.
Wireless makes it easy and cost-effective to get better insight
into what’s happening, especially for safety and security. For exam-
ple, it’s easy and cost-effective to add wireless cameras where it
would be too difficult, costly or risky to run wires. Greg Martin is senior process control consultant with Emerson
Process Management. He has 30 years’ experience in automation
Wireless closed-circuit television cameras and RFID-equipped consulting, specializing in advanced process control. Dr. Martin has
access badges also enable intelligent security monitoring and BS and MS degrees from Oklahoma State University, a PhD degree
control—from restricting access to specific areas based on levels from Purdue University and is a registered professional engineer in
of security, to tracking attempts to violate security protocols, to Texas. He has authored 60 publications and holds nine patents.
32
OPC UA: an end user’s perspective

The updated specification relies on Web services for its data
transportation providing significant advantages
R. KONDOR, OPC Training Institute, Edmonton, Alberta, Canada
O
PC UA (unified architecture) represents the OPC Foun- OPC enables plants to automate the transfer of data from a con-
dation’s most recent set of specifications for process control system (PLC, DCS, analyzer, etc.) to an industrial software
trol and automation system interconnectivity. This article application (HMI, historian, production system, management
explains OPC UA from the perspective of the organization that system, etc.). OPC is typically found in Level 3 networks and
will benefit from the connectivity—the end user. higher. Thus, OPC transfers process control data between the con-
The first form of OPC relied on DCOM for its data trans- trol (Level 2) network and the operations/manufacturing (Level
portation, which was very powerful and versatile, but posed a 3) network. It also exchanges data between the operations/manu-
problem for those who did not understand how to configure facturing network and the business (Level 4) network. In essence,
DCOM. Instead of DCOM, OPC UA relies on Web services for OPC is the Modbus of the new century. It is not a replacement
its data transportation. OPC UA also uses objects to help with for low-level communication standards such as 4-20mA, HART,
data description. Even though these are major additions and Profibus, or Foundation fieldbus. Rather, organizations use OPC
modifications to OPC, OPC UA will be backward compatible in high-level communication.
with older products through the use of wrappers. All this will Note: OPC is no longer an acronym. When OPC first released
ensure that OPC UA will be even better suited to penetrate the in 1996 it was an acronym for OLE for process control, and was
entire plant enterprise. Of course, with all the new connectivity restricted to the Windows operating system. OPC is now avail-
that OPC UA offers, the new challenge will be system security. able on other operating systems and enjoys significant adoption
outside of process control. So, the original name is no longer
OPC overview. OPC is an industrial communication standard appropriate and OPC changed from an acronym to a word.
that enables manufacturers to use data to optimize production,
make operation decisions quickly and generate reports (Fig. 1). OPC communication started with DCOM. When
OPC was first released in 1996, with the OPC Data Access 1.0
(OPC DA 1.0) specification, it used Microsoft’s DCOM as the
data transportation mechanism (Fig. 2). Data moved between
OPC applications on different computers using DCOM. At the
time, DCOM was an outstanding choice because it provided a
working communication infrastructure complete with all the
necessary security services (authentication, authorization and
encryption). Thousands of vendors were already using DCOM
because it was a relatively versatile application programming
interface (API). DCOM was a clear winner at the time, but
while it provided a reliable communication backbone for OPC,
it did have several challenges.
First, DCOM configuration eludes automation personnel who
do not take time to learn it. DCOM is actually very predictable
and is not difficult to configure. While there are training classes
that explain DCOM configuration in detail, most people do not
take the time to learn it and so DCOM’s behavior frustrates them.
Consequently, automation personnel needlessly experience prob-
lems when connecting two computers and configuring firewalls.
Nevertheless, knowledgeable users can easily configure DCOM
in a matter of minutes.
Second, many programmers assume that network communi-
cation will occur without any data loss. This assumption leads
them to create products that are highly susceptible to data loss
FIG. 1 OPC communication enables applications to interoperate and communication timeouts. As a result, end users might some-
and simplifies system architecture. times experience a delay in application responses and complain to
I 33
their vendors. However, since the programmers fail to understand etc.) typically requires a PC anyway. Nevertheless, it would be
DCOM’s behavior, they often incorrectly blame DCOM for poor possible to have a PLC communicate with a software application
application behavior, which further promotes the false myth that using OPC without requiring an intermediate computer that
DCOM is unreliable. Again, informed programmers are easily uses Windows.
able to compensate for data loss and are able to make DCOM
work reliably, and in a way that end users would expect. OPC UA uses an object-oriented data model. Classic
The third problem is DCOM does not work through network OPC has a fairly simple data model. Each of the OPC specifica-
address translation (NAT). Thus, DCOM does not work in the tions handles a different aspect of the data. For example, the OPC
rare cases where communication must occur between two private DA (data access) specification communicates real-time values, the
networks that are separated by a public network. Such is the case OPC HDA (historical data access) specification communicates
when two plants attempt communication over the Internet. NAT archived values, the OPC A&E (alarms and events) specification
is sometimes used inside industrial facilities, but this is often communicates various process and system events (such as a tem-
unnecessary since a firewall would suffice. perature that exceeds a prespecified limit), and so on. In addition,
The fourth problem is DCOM is proprietary to Microsoft. classic OPC implements each specification separately; essentially
This makes OPC difficult (to impossible) for vendors to port to in a different executable. Thus, it is time-consuming to match
nonWindows operating systems. While some vendors are able to item names with real-time data and historical data. Even worse,
embed Windows directly on their own controller (PLC, DCS, automated applications may not be able to do it at all.
analyzer, etc.) hardware, others are unable to do this. Also, com- OPC UA provides a unified data model. Thus, when an appli-
panies that use nonWindows operating systems (UNIX/Linux, cation uses OPC UA to send a temperature reading, the receiver
VMS, etc.) are having a difficult time importing OPC data into is able to retrieve the real-time value, any associated historical
their applications. values, and even alarms and events. All these data are available
from pointing at a single OPC item. The OPC server is able to
OPC UA uses Web services. OPC UA uses Web services associate all the data together so that the OPC client does not
instead of DCOM for data transportation. This change most need to redo the association work. For example, in DCOM-
end users will notice immediately. Two of the biggest advantages based OPC, end users who are interested in a pressure reading
of Web services are ease of communication would have had to point to the OPC DA
between networks and independence from server to look at the real-time value. Then
specific operating systems. The challenge they would have to point to an OPC HDA
for the plant will be implementing security server to trend the pressure over the past
to keep the data safe. shift. If they wanted to take a look at asso-
Perhaps the biggest technical advantage ciated events, they would have to point to
of Web services is that they enable OPC the OPC A&E server. But with OPC UA,
to communicate over a single port using the end user can simply point to a pressure
a protocol that most firewalls will allow reading, view its real-time value, look at the
to pass by default. This should make it past shift’s trend (historical data) and view
easier for integrators to set up a system for all the associated events by connecting to a
communication between networks. Many single OPC UA server.
firewalls are already configured to let Web OPC UA also provides the ability to cre-
traffic pass across port 80. This will make ate more complex objects. For example, one
it easier for IT to open the ports necessary could create a pump that is composed of
to implement OPC communication. Previ- various temperature, level, pressure, flow
ously, DCOM required multiple ports to and vibration readings. Included would
establish communication. While this was be the history of all values as well as a pic-
possible to configure, a significant portion ture of the pump. One could even associ-
of automation personnel did not take the ate P&ID schematic diagrams and main-
time to learn how to do it. Nevertheless, tenance orders. This presents a powerful
opening port 80 opens communication for mechanism for integrators from various
a plethora of applications (not just those companies to share data without having
that are needed for operations), so emphasis to recreate it in their different proprietary
on security will be required immediately. software applications.
In addition, Web services are not bound
to any specific operating system. Thus, ven- Improving existing specifications.
dors will have an easier time implementing As OPC evolved over the years, the OPC
OPC servers on their automation hard- Foundation provided constant updates and
ware and nonWindows operating systems. improvements to the specifications. OPC
Vendors are already working on PLCs that UA continues this tradition. After con-
include an embedded native OPC server sulting end users, integrators and vendors,
that does not require an external computer. the OPC Foundation decided on various
However, this implementation might not additions to the specifications to handle
be as simple as it seems because an auto- FIG. 2 OPC initially relied on DCOM for the most common challenges. OPC UA
mation application (HMI, historian, APC, data transportation. includes mechanisms to quickly inform
34
users of broken communication, identify lost data and even that any new OPC UA product will communicate with an existing
provide for redundancy. DCOM-based OPC product. As a result, there is no need to con-
OPC UA uses a poll-report-by-exception mechanism. Thus, template whether or not one should wait for OPC UA products.
the OPC client polls the OPC server for changes. The server It is easy to implement DCOM-based OPC products today and
then responds with any data changes. A failure to respond would be assured that future OPC UA products will communicate with
immediately tell the OPC client that the communication is no the old software.
longer active. In addition, updates can come as quickly as the Two wrappers will be available: one for OPC clients and the
polling itself. However, unlike common protocols that must other for OPC servers. The first wrapper will convert a DCOM-
poll each point individually and consume precious bandwidth, based OPC server to an OPC UA server. Thus, an OPC UA
OPC UA enables the OPC server to respond with any data that client will be able to connect to the existing DCOM-based OPC
changed. Thus, a single efficient poll can bring back a large server without any changes. The second wrapper will convert a
amount of data that include all the changes in the process as DCOM-based OPC client to an OPC UA client. So an existing
well as the health of the OPC server itself. By contrast, before DCOM-based OPC client application (such as an HMI) will
OPC UA, DCOM communication sent all changes to the OPC be able to communicate with an OPC UA server that could be
client by exception. Thus, an OPC client did not have to poll purchased a year from now. Using wrappers, OPC is sure to ease
the OPC server periodically. While this was efficient, many the transition from the old to the new technology.
programmers overlooked the possibility that no updates would Wrappers will tunnel OPC to places where DCOM-based
be received when communication breaks. As a result, the OPC OPC can’t penetrate on its own. For example, when an OPC
client would wait for updates that would never arrive. Various client and server are separated by NAT, DCOM will not be able
companies overcame these difficulties, but some did not and to make the connection. However, by converting the DCOM-
blamed DCOM instead. based call to OPC UA at the source, and converting it back from
OPC UA also enables an easier implementation of redun- OPC UA to DCOM at the destination, the call will transport the
dancy. OPC UA servers can update a set of clients. By contrast, required data. Tunneling will likely be the first form of OPC UA
DCOM-based OPC servers could only update OPC clients that implementation as OPC UA products begin to emerge.
explicitly subscribed to the data. As well, since the OPC client
can easily tell when communication with an OPC server fails Shop floor to top floor: OPC to the enterprise. OPC
(as above), the OPC client can now quickly failover to a standby UA introduces an object model to industrial data, and Web services
OPC server. In DCOM-based implementations, most vendors will enable the OPC applications to transport the data across fire-
relied on third-party OPC redundancy applications that cost walls, networks and the Internet (Fig. 4). A variety of applications
them additional funds.
M3 Technology
M3
Backward compatibility and tunneling. The OPC
Foundation has promised to supply the industry with two simple
software applications that will enable people to quickly convert
their DCOM-based OPC products to OPC UA. These software SIMTO™
applications are called “wrappers” (Fig. 3). Wrappers will ensure Advanced Scheduling, Planning
and Optimization Solutions
www.m3tch.com
X SIMTO Scheduling
Oil Refining
Petrochemical
LNG (liquefaction & regasification)
Terminals
X SIMTO M-Blend
Multi blend recipe optimization
Gasoline, Crude, Fuels, Asphalt
Naphtha olefin plant feedstock
X SIMTO Dock Manager
Jetty/berth scheduling
X SIMTO Distribution
Supply & distribution optimization
X SIMTO Planning Workspace
Sales and operation planning
10850 Richmond Ave., Suite 290, Houston, TX 77042
FIG. 3 OPC UA wrappers will enable legacy DCOM-based OPC
products to communicate with new OPC UA products. Tel: +1.713.784.8285 • Fax: +1.832.553.1893
35
S T E A M
U T I L I TY
SOLUTIONS
http://www.armstronginternational.com/HPI
In the face of rising energy costs, let Armstrong optimize your facility’s steam utility system.
For more than 100 years, Armstrong International has provided utility optimization for our global partners.
We’ve solved problems, conserved energy and improved efficiency in countless applications.
We can do the same for your petrochemical facility with our complete prefabricated piping solutions. Designed
to simplify and supply all the components necessary for your drip and tracer line applications, Armstrong’s steam
distribution manifolds, condensate collection manifolds and trap valve stations bring everything together. You’ll
enjoy lower installation costs and a compact, easy-to-access, centrally located assembly.
We also offer complete steam system asset management. Our professionals can conduct trap audits, deliver
a system analysis and recommend ways you can optimize. To reduce energy costs while ensuring best-of-class
performance, contact your Armstrong representative or visit armstronginternational.com/HPI. © 2008 Armstrong International, Inc.
Select 65
Select 68 at
at www.HydrocarbonProcessing.com/RS
INSTRUMENTS AND NETWORKS
Cut field connections

and potential leak
points – tenfold.
FIG. 4 OPC UA will enable data to move from the shop floor to
the top floor.
will be able to supply the enterprise with data. An HMI will be able
to pass equipment events to the maintenance system. The historian
will be able to pass calculations to various engineering systems. As
well, inventory management systems will be easily able to obtain
production figures directly from automation equipment.
Plant-floor data will finally find its way to the business local
area network (LAN) and enable a variety of applications to benefit
from the newly available data. For instance, computer mainte-
nance management systems (CMMSs) or enterprise asset manage-
ment systems (EAMSs) will be able to obtain equipment condi-
tion data so they can implement a conditions-based maintenance
(CBM) program. Enterprise resource planning (ERP) applications
will be able to obtain inventory information, or even send produc-
tion orders without any manual intervention.
Security: the new challenge for automation. OPC

UA makes it relatively easy for a multitude of applications to con-
nect with each other. So the new challenge for automation person-
nel will be to secure their systems from unwanted connections.
Web services will make it easy to cross firewalls and networks. So, With Armstrong’s compact manifold
unwanted connections from people and applications will become system for steam distribution and
far more common. However, unlike IT systems, automation sys- condensate collection.
tems are responsible for production and safety. Therefore, security
will quickly rise to the forefront. Automation personnel will have Armstrong’s modular steam tracing systems will:
to learn how to secure their systems in a way that still enables • Lower your installation costs
them to provide access to those who need it. • Reduce time spent in design and construction
It remains to be seen how vendors will enable their applica- • Lower long-term maintenance and operating costs
tions with the three pillars of secure connectivity: authentication, • Provide advanced piston sealing technology,
authorization and encryption. Various products that are already reducing overall life-cycle costs
in the planning stages still do not include the necessary facilities
for proper security. These applications use “security by obscurity,” Contact your Armstrong representative or visit
which essentially relies on a hacker’s inability to understand how a armstronginternational.com/HPI.
system works to make it behave inappropriately. Both process and
attitudes toward security will have to change. HP
Randy Kondor is a computer engineer and president of the

OPC Training Institute, the world’s largest OPC training company.
Since 1996, he has been vastly involved within the OPC industry © 2008 Armstrong International, Inc.
and a strong supporter of the OPC Foundation. Mr. Kondor contin-
ues to dedicate himself to spreading the OPC Foundation’s message
about system interoperability and intervendor cooperation.

37
You Get More Than Just a Process Gas Compressor
Lubricated up to 1’000 bar, non-lubricated up to 300 bar
For longest running time: We recommend our own designed,

in-house engineered compressor valves and key components
Designed for easy maintenance
We are the competent partner with the full range

of services – worldwide
Your Benefit:
Lowest Life Cycle Costs
More benefits: www.recip.com/api618

Soft sensor modeling using

artificial neural networks
Here are guidelines for proper construction
V. NANDAKUMAR, Mangalore Refinery & Petrochemicals Ltd., Karnataka, India
W
ith increased competition and rising feedstock cost, • Implementation in real time
pressures are increasing for refinery managers to extract • Periodic data validation and tuning.
maximum value out of processes. In operations, the To elucidate the concepts, a real example of a CCR-platform-
online quality monitoring is an important part of process control. ing process unit and the product’s research octane number (RON)
Typically, analyzers are provided for this application. However, as as required product quality to be monitored was taken.
parameters increase in complexity from density, moisture content,
octane number, sulfur content, etc., the cost and maintenance CCR process brief. The process produces feed for an aromatics
efforts on analyzers increase exponentially. Moreover, the inherent complex or a high-octane gasoline blending product and a signifi-
rigidity in hardwired analyzers makes their cant hydrogen as a byproduct. In the unit,
extended usage difficult, if not impossible, ■ Basic requirements for a hydrotreated naphtha feed is combined with
so soft sensors play a vital role. recycle hydrogen gas and heat exchanged
Soft sensors have the advantages of easy soft sensor are the knowledge against reactor effluent. The combined
maintainability, low cost and extensibility feed is raised to reaction temperature in the
to other applications. It’s not difficult to of fundamental relationships charge heater and sent to the reactor section.
design a soft sensor for new parameters Radial-flow reactors are arranged in a vertical
and one can be virtually built for every of process variables and the stack. The predominant reactions are endo-
parameter in question. Basic requirements parameter in question. In thermic, so an inter-heater is used between
for a soft sensor are the knowledge of fun- each reactor to reheat the charge to reac-
damental relationships of process variables short, it is a sophisticated tion temperature. The effluent from the last
and the parameter in question. In short, it reactor is heat exchanged against combined
is a sophisticated correlation model. correlation model. feed, cooled and split into vapor and liquid
In petroleum refining, correlations and products in a separator. The vapor phase is
empirical relations have played an important historical role in rich in hydrogen gas. A gas portion is compressed and recycled back
plant design and operations. The advent of inexpensive comput- to the reactors. The net hydrogen-rich gas is compressed and purified
ing power made direct computation models like finite element in a PSA system. Catalyst flows vertically, by gravity down the reactor
analysis (FEA) and computational fluid dynamics (CFD) feasible stack. Over time, coke builds up on the catalyst at reaction condi-
and practical. The skill set requirement for such applications tions. Partially deactivated catalyst is continually withdrawn from the
allows for highly trained experts to test. In refineries and produc- reactor stack bottom and transferred to the CCR regenerator.
tion departments, the required skills are different; hence, a CFD
or FEA may not be feasible in day-to-day applications. Empirical Variable identification. A reformate product is an impor-
relations find the best acceptance in quick calculations and the tant component in gasoline blending as well as feed stock for the
not-so-accurate control algorithms. API data books list many downstream mixed xylene unit. The key parameter for this prod-
empirical relations between properties of hydrocarbon liquids uct is its RON. Online analyzers are usually provided in the unit.
such as relations between density and boiling point, molecular However, most often these analyzers have maintenance problems
weight, flash point and initial boiling point relations. These rela- and require frequent offline calibrations that require laboratory
tions place less importance on the underlying theoretical models analysis data.
than the accuracy of the results. It is more like a “black box” The process technology manual of the licensor suggests that
approach. the product RON is a function of the following variables:
Soft sensor construction can be split into various steps as fol- • Feed rate (analogously the liquid hourly space velocity
lows: (LHSV) or residence time)
• Variable identification • Feed quality—described by naphthenic and aromatic content
• Data collection • Reactor severity
• Programming • Hydrogen partial pressure
• Sensor testing • Catalyst activity
I 39
However, all of the above cannot and are not directly measured
by instrumentation; hence, a soft sensor, though accurate and Hidden layer
constructed with those variables, is impractical to use. Therefore,
process “proxies” are suggested that are monitored online and
easy to configure. Input layer
The following variables are substituted:
Output layer
• Feed quality by reactor total temperature difference which is
the weighted sum of each individual reactor delta and reformate
product flow
• Hydrogen partial pressure by a total reactor pressure, recycle
gas flow, net gas flow
• Catalyst activity can be substituted with a catalyst circula-
tion rate with the coke deposition on a catalyst which in turn
was substituted by total air demand and regenerator peak burn
temperature. In this particular example, the actual data of catalyst
circulation and regenerator variables did not vary; hence, their
effect was constant toward RON and was not considered. FIG. 1 Standard FFnet.
Data collection. In the operating unit studied, the daily

reformate sample goes to the laboratory at 7 am. Ideally, the response for the example set. The resulting network should then
RON result is the net effect of that particular instant which be able to generalize (give a good response) when presented with
is enveloped in the previous three hours of operation due to cases not found in the set of examples. Unsupervised learning is an
various hold-ups and residence time in the system. Due to the automated process, but details of the process are omitted here.
difficulty in collecting precisely enveloped data from the plant A popular neural net structure is the feed-forward neural net-
historical database, it was later approximated with the daily work (FFnet). They are known by another name, multi-layer
average values. perceptrons. In a feed-forward neural network, the neurons are
Statistically, a correlation is useful if it is taken with strictly usually arranged in layers. A typical layer structure is:
independent variables. To ensure that, from the collected data, Input t Layer 1 t Layer 2 t Output
paired covariance analysis was conducted and the group was Layers 1 and 2 are the inner layers and are labeled as hidden.
confirmed to be reasonably independent. Since it was possible to The connections are always forward, e.g., from Layer 1, every neu-
have a certain degree of parametric relations between the variables, ron connection is to Layer 2 or to an output layer only. An FFnet
a judgmental decision was taken. has neurons arranged in a distinct layered topology. The input layer
The following values were collected and examined for any is not really neural at all—these units simply serve to introduce
abnormality such as the instrument showing too high or too the input variable values. The hidden and output layer neurons
low values, missing data, etc., and then corrections were made are each connected to all units in the preceding layer. Although it
wherever necessary. is possible to define networks that are partially connected to only
Independent variables (input variables): some units in the preceding layer, for most applications, fully con-
• Unit feed rate nected networks are used. In a standard FFnet, the connections are
• Reactor severity strictly to the next layer and all nodes are connected to the next
• Reactor total delta temperature layer nodes. A typical standard FFnet is illustrated in Fig. 1.
• Reformate rate Each arrow in the figure symbolizes a parameter in the net-
• Reactor pressure work. The network is divided into layers. The input layer consists
• Recycle gas flow, and of network inputs and then follows a hidden layer, which consists
• Net gas flow of any number of neurons, or hidden units placed in parallel. Each
Dependent variable (output variable): neuron performs a weighted summation of the inputs, and then
• Reformate RON as reported by the lab. passes a nonlinear activation function, F. The network output is
formed by another weighted summation of the outputs of the
Feed-forward artificial neural networks. A neural neurons in the hidden layer.
network is an information processing structure consisting of The network calculations are progressively applied with input
processing elements (neurons), interconnected with directional layers simply taking the values of input vectors. Each hidden and
signal channels called connections. Each processing element has output layer is calculated by the activation value by taking the
a single output connection that branches into as many collateral weighted sum of the outputs of the units in the preceding layer,
connections as desired. Each carries the same signal—the pro- and subtracting the threshold value. The activation value is passed
cessing element output signal, which can be any mathematical through the activation function to produce the neuron output.
type desired. Neural networks develop information processing When the entire network has been calculated, the outputs of the
capabilities by learning from examples. Learning techniques can output layer act as the output of the entire network.
be roughly divided into two categories: supervised and unsuper- A widely used activation function is the sigmoid function
vised learning. given by:
Supervised learning requires an example set where the desired 1
network response is known. The learning process consists in F (x ) = (1)
adapting the network in a way that it will produce the correct 1+ e x
40
The net input to a processing unit, h, is given by:

Input layer
net h = wth xi + j (2)

i
where xi s are the outputs from the previous layer, wih is the weight
(connection strength) of the link connecting unit i to unit j, and ␪
is the bias of unit h, which determines the location of the sigmoid
function on the x-axis.
The activation value (output) of unit j is given by:
1
ah = F (net h ) = (3)
net
1+ e h
The objective of different supervised learning algorithms is the
iterative optimization of a so-called error function representing a
measure network performance. This error function is defined as
the mean square sum of differences between the output unit values
of the network and the desired target values, calculated for the
whole pattern set. The error for a pattern p is given by:
NO
E p = (d pj a pj )2 (4)
j =1
where dpj and apj are the target and the actual response value of Hidden layers
output neuron j corresponding to the pattern p. This factor is
improved successively by the learning algorithms and the model
is set when the total error is minimum.
The total error is: Output layer
NO
P
1 1 P
E = E p = (d pj a pj )2 (5)
p=1 2 2 p=1 j =1 FIG. 2 CCR unit RON model as FFnet {7,[7,7],1}.
where p is the number of the training patterns.

The actual computation process in the FFnet is quite involved This choice was albeit arbitrary, and has medium complexity. In
and not practical to do manually. The software packages specifi- theory, it is a trade-off between accuracy and computation. The
cally designed for this application are available in both commercial selected FFnet is described as FFnet {7, [7, 7], 1}. The square
and free open source domains. bracket denotes the hidden layers. The network is illustrated in
Typically, these applications require the user to specify the Fig. 2. The program was written in python and the data, which
network topology, input vector and target vector. The inner work- were daily average input and output value vectors for 249 days.
ings are conveniently shielded from user interface. Of course, the The program does a statistical testing to find the neural net regres-
open-source version does permit such changes in the program, sion fit. The data set consists of daily averages with seven param-
but, in a normal course such modifications are not needed. eters and the target RON values for training.
In the above case, the entire program is scripted in Python, an
open source and a highly powerful programming language. The Python code for fitting and training the data for
FFnet code was provided by a library module called ffnet. The FFnet:3
details of the FFnet structure are: # importing the required modules
• Network has feed-forward architecture import win32com.client
• Input units have identity activation function and all other from ffnet import mlgraph,ffnet, savenet, exportnet
units have sigmoid activation function import pylab as p
• Provided data are automatically normalized, both input # use psyco to speed up
and output, with a linear mapping to the range (0.15, 0.85). import psyco
Each input and output is treated separately (i.e., the linear map is psyco.full()
unique for each input and output). # data is read from the file “ccr1 data new.xls”
• Function minimized during training is a sum of squared xlApp=win32com.client.gencache.EnsureDispatch(“Excel.
errors of each output for each training pattern. Application”)
The module has an added feature that the trained neural net xlWb=xlApp.Workbooks.Open(“C:\Users\Admin\Documents\
can be exported as a FORTRAN routine that can be compiled to ccr1 data new.xls”)
use in other systems. xlSht=xlWb.Worksheets(1)
row_range=range(4,253)
Model implementation. As mentioned in the data collec- dtemp=[]
tion section, the input vector has a dimension of 7!1 and the wait=[]
output vector is 1!1. feed=[]
The FFnet selected has seven input nodes and one output reformate=[]
node. Two hidden layers, each having seven nodes, was chosen. pressure=[]
I 41
netgas=[]
Average percent error in RON fit using ANN
rggas=[] 1.0
RON=[]
points=len(row_range)
for i in row_range: 0.5
dtemp.append(float(xlSht.Cells(i,3).Value))
wait.append(float(xlSht.Cells(i,4).Value))
APE, %
feed.append(float(xlSht.Cells(i,5).Value)) 0.0
reformate.append(float((xlSht.Cells(i,6).Value)))
pressure.append(float((xlSht.Cells(i,7).Value)))
netgas.append(float((xlSht.Cells(i,11).Value)))
–0.5
rggas.append(float((xlSht.Cells(i,15).Value)))
RON.append(float((xlSht.Cells(i,18).Value)))
# input data set includes feed, wait, dtemp, pressure, reformate, –1.0
rggas, netgas 0 50 100 150 200 250
data_set=[]
FIG. 3 Result of model fit showing APE.
for n in range(points):
data_set.append([feed[n],wait[n],dtemp[n],pressure[n],
reformate[n],\ rggas[n],netgas[n]]) net.train_rprop(input, target, a=1.2, b=0.5,\
xlApp.ActiveWorkbook.Close(SaveChanges=0) mimin=9.9999999999999995e-07, mimax=50.0, xmi=0.10,
xlApp.Quit() maxiter=10000,\ disp=1)
# defining ffnet parameters print “TRAINING NETWORK...”
input = data_set net.train_tnc(input, target, maxfun = 5000, messages=1)
target= RON # Test network
# making a network print
conec=mlgraph((7,7,7,1)) print “TESTING NETWORK...”
net=ffnet(conec) output, regression = net.test(input, target, iprint = 2,\
# using resilient propagation algorithm filename=“ccr_yearly_test.txt”)

42
TABLE 1. Model’s average percent error for use in plant operations. The sensors will also help fine-tune the
existing APC systems. HP
No. RON Lab data RON prediction APE
1 100.9 99.0 –1.9% LITERATURE CITED
1 UOP Brochure on CCR-Platforming was referred for the process descrip-
2 100.4 100.3 –0.1% tion.
2 Barbălată, C. and L. Leustean, “Average monthly liquid flow forecasting using
3 100.7 99.0 –1.7%
neural networks.”
4 100.1 99.8 –0.3% 3 www.python.org
5 100.8 97.1 –3.6% 4 Wojciechowski, M., “Feed-forward neural network for python,” [FFNET,
6 101.3 99.2 –2.0% 2007]{FFNET}, Technical University of Lodz (Poland), Department of Civil
Engineering, Architecture and Environmental Engineering, http://ffnet.
7 101.3 99.3 –1.9% sourceforge.net, ffnet-0.6, March 2007.
8 100.7 99.5 –1.2%
AUTHOR’S NOTE
9 100.4 98.6 –1.8%
The author would like to thank Ms. Lakshmi T. N. V. for being a contributing
10 100.4 100.5 0.1% author. She is a chemical engineering graduate and worked as a process engineer
11 100.5 100.6 0.1% with the Mangalore Refinery and Petrochemical Limited (MRPL). Ms. Lakshmi’s
assistance extended in retrieving and analyzing field data from DCS and model
building, and then presenting the results.
# Exporting network
savenet(net, “ccr_yearly”)
# exporting the net as a FORTRAN module to use later V. Nandakumar is a senior technical manager at the Man-
exportnet(net, “ccr_yearly.f ”) galore Refinery and Petrochemical Limited (MRPL), a subsidiary of
Oil and Natural Gas Corporation Limited. His current assignments
# Plotting the data include appraisal of new project plans, plant configurations and
RON_fit=[] frontier technology analysis in refining processes for review by
for n in range(points): upper management. Mr. Nandakumar has over 15 years of operational experience
RON_fit.append(net(input[n])) with secondary processing units including naphtha reformers, sulfur recovery pro-
cesses, the operation and commissioning of CCR-platforming units, process design,
# Average Percent Error APE HAZOP analysis, and quality and environmental management program implementa-
APE=[] tion under ISO standards. He has a special interest in the application of IT tools in
for n in range(points): chemical engineering, mostly open-source software. Mr. Nandakumar received his
APE.append((RON[n]–RON_fit[n][0])*100/RON[n]) BTech degree in chemical engineering from the University of Calicut, Kerala.
p.plot(range(points),APE,’b’)
p.title(“Average Percent Error in RON fit using ANN”)
p.ylabel(“APE %”) INSTRUMENTS and
p.grid(True)
p.show()
MAINTENANCE
Where do you stand?
The regression result is given below:
Feed-forward neural network:
Inputs: 7
Hidden: 14
Outputs: 1
Connections and biases: 120
Testing results for 249 testing cases: This extensive 12-part DVD series covers many aspects
pects
ect off mechanical
mecha
OUTPUT 1 (node nr 22): maintenance and provides a long needed source of practical
Regression line parameters: engineering reference information that the viewer can easily adapt
Slope = 0.910627 to similar machinery or machinery installations in a particular plant.
Intercept = 9.021174 Visit www.GulfPub.com/MachineryCompSeries
Correlation = 0.953035
Tail probability = 0.000000
Standard error = 0.259538
The error pattern after training is illustrated in Fig. 3 for the

249 data points. UPDATED VERSION FOR 2009!
After fitting and training the FFnet model, it was tested with Software forr inst
instrumentation
trumentation design and selection
selection, sizing more than
a different set of data and the accuracy is tabulated and listed in 50 different instruments. This is the only sizing program you need
Table 1. to consider!
The FORTRAN module exported from the Python program can Visit www.GulfPub.com/InstruCalc
be compiled with the standard FORTRAN compilers with the stan-
dard link file from the FFfnet module (given with the installation of Gulf Publishing Company
the module), and can be used for any APC implementation. +1 713 520 4428 l +1-800-231-6275
+1-713-520-4428 +1 80
00 231 627
75
The above example shows the ease and availability of standard Email: [email protected]
software tools to model software sensors with reasonable accuracy
43
KINETICS TECHNOLOGY
INTERNATIONAL
THE WORLD LEADER

E N G I N E E R I N G – F A B R I C AT I O N
C O N S T R U C T I O N – N Ox R E D U C T I O N
F I R E D H E AT E R S
Refinery Applications
Steam Reformers
Petrochemical Applications
OTSGs
NOx/CO REDUCTION
K T I C O R P O R AT I O N
Gas Turbines
1990 Post Oak Blvd., Suite 200, Houston, TX 77056
Heaters
Tel: (281) 249-2400 Fax: (281) 249-2328
Boilers E-mail: [email protected]
FCC Units KTI - KOREA
other fired sources #612, Kolon Science Valley II, 811, Guro-dong,
Guro-gu, Seoul, 152-050, Korea
Please visit www.kticorp.com for a complete Tel: 82-2-850-7800 Fax: 82-2-850-7828
list of our products, services, and contacts. E-mail: [email protected]
Hydrogen gas detection

Combining detection systems improves safety
E. NARANJO, General Monitors, Lake Forest, California
O
il refineries are large hydrogen gas producers and con- In oil refineries, the first step in fire escalation and detonation is
sumers. Hydrogen plays a pivotal role in many refining loss of containing the gas. Hydrogen leaks are typically caused by
operations, from hydrocracking—heavy gas reduction defective seals, valve misalignment, or flange or other equipment fail-
and gasoils to lower molecular weight components—to gas stream ure. Once released, hydrogen diffuses rapidly. If the leak takes place
treatment, to catalytic reforming. In catalytic reforming, the gas outside, the cloud dispersion is affected by wind speed and direc-
is also used to prevent carbon from reacting with the catalyst to tion, and can be influenced by atmospheric turbulence and nearby
maintain the production of lighter hydrocarbons while extending structures. If the gas is dispersed in a plume, a detonation can occur
the catalyst’s life. Not surprisingly, refineries use large volumes if the hydrogen and air mixture are within its explosion range and
of hydrogen that is either produced onsite or purchased from an appropriate ignition source is available. Such flammable mixtures
hydrogen production facilities. can form at a considerable distance from the leak source.
Demand for hydrogen is growing. Changes in gasoline and To address the hazards posed by hydrogen, fire and gas detec-
diesel fuel specifications, prompted by environmental legislation system manufacturers work within the construct of protec-
tion, have led to increased hydrogen use to improve gasoline tion layers to reduce hazard propagation incidences. Under such
grade. However, higher crude oil prices have enhanced the com- a model, each layer acts as a safeguard, preventing the hazard from
mercial prospects of heavier crudes, requiring new investments becoming more severe. Fig. 1 illustrates a hazard propagation
in conversion processes and more extensive hydrotreating and sequence for hydrogen gas leaks.
hydrocracking applications. Detection layers encompass different sensing techniques that
The scale and growth of hydrogen demand raises the fundamen- either improve scenario coverage or increase the likelihood that a
tal question about using the gas safely. Due to its chemical proper- specific type of hazard is identified. Such fire and gas detection layers
ties, hydrogen poses unique challenges in the plant environment. can consist of catalytic sensors, ultrasonic gas leak monitors or fire
Hydrogen gas is colorless, odorless and undetectable by human detectors, which are illustrated in Fig. 2. Ultrasonic gas leak detectors
senses. Also, hydrogen is not detected by infrared (IR) gas-sensing can respond to high-pressure releases of hydrogen, such as those that
technology. Since it is lighter than air, it is difficult to detect where may occur in hydrocracking reactors or hydrogen separators. Con-
accumulations should not occur. Coupled with the challenge of gas tinuous hydrogen monitors, like catalytic detectors, can contribute
detection are the safety risks posed by the gas itself. to detecting small leaks. Leaks may happen when a flange slowly
A practical approach is offered for the deployment of fire and deforms by use or failure of a vessel maintained at or near atmo-
gas detectors that maximize detection efficiency. The approach is spheric pressure. To further protect a plant against fires, hydrogen-
that any one detection technique cannot respond to all hazardous specific flame detectors can supervise entire process areas. Such wide
events. Consequently, the risk of detection failure is reduced by coverage is necessary since a fire may ignite at a considerable distance
deploying devices that have different strengths and limitations. from the leak source due to hydrogen cloud movement.
When a containment system fails, hydrogen gas escapes at a
Improved safety through diversity. There are several rate that is proportional to the orifice size and the system’s internal
hazards associated with hydrogen that include: respiratory ail- pressure. Such leaks can be detected by ultrasonic monitors that
ment, component failure, ignition and burning. Although hazard sense airborne ultrasound produced by turbulent flow above a
combinations occur in most instances, the primary hazard with pre-defined sound pressure level. Using ultrasound as a proxy for
hydrogen is a flammable mixture production that can lead to a fire gas concentration is a major technique advantage. Ultrasonic gas
or an explosion. Hydrogen is easily ignited since the minimum leak detectors do not require gas transport to the sensor element
ignition energy at atmospheric pressure is about 0.2 mJ. to detect gas. They are unaffected by leak orientation, gas plume
In addition to these hazards, hydrogen can produce mechani- concentration gradient and wind direction. Such features make
cal failures of containment vessels, piping and other components
due to hydrogen embrittlement. Metals and plastics can lose
ductility and strength due to long-term exposure to the gas. This Equipment Gas Property damage/
leads to crack formation and eventually causes ruptures. A form Ignition Fire/explosion
rupture dispersal personal injury
of hydrogen embrittlement takes place by a chemical reaction.
At high temperatures, hydrogen reacts with one or more metal-
wall components to form hydrides that will weaken the material FIG. 1 Hazard sequence for hydrogen dispersal. Layers of
lattice structure. protection separate each hazard state.

I 45
110
Fire/explosion protection 105
100
Fire detection
95
SPL, dB
90
Ventilation
85
80
Gas detection
75
70
0 2 4 6 8 1 01 21 4
Leak detection Distance from source, m
FIG. 3 Sound pressure level as a function of distance for

hydrogen leaks. Leak size = 1 mm diameter orifice,
differential pressure = 5,515 kPa (800 psi), leak rate =
Containment 0.003 kg/s.
but are generally tailored for maximum selectivity and response

FIG. 2 Protective barrier schematic for a hydrogen accident speed. As pointed out earlier, hydrogen cannot be detected by IR
sequence. absorption, making catalytic monitors one of the most reliable
technologies for hydrogen gas detection.
ultrasonic gas leak detectors an ideal choice for the supervision of Along with catalytic and ultrasonic gas leak detectors, hydrogen-
pressurized pipes and vessels in open, well-ventilated areas. specific flame detectors add another barrier against the propagation
Another instrument advantage is the wide coverage area per of hydrogen hazards. The instruments simultaneously monitor IR
device. Depending on the ultrasound background level, a single and ultraviolet (UV) radiation at different wavelengths. Radiation
detector can respond to a small hydrogen leak at about 8 m from the is emitted in the IR by water molecules created by hydrogen com-
source. As illustrated in Fig. 3, even small leaks can generate sufficient bustion. The emission from heated water or steam is monitored in
ultrasonic noise to afford detection in most industrial environments, the wavelength span from 2.7 ␮m to 3.2 ␮m. An algorithm that
where background noise levels can range from roughly 60 dB to 90 processes the modulation of IR radiation allows the detectors to
dB. Since the instrument responds to the gas release rather than the avoid false signals caused by hot objects and solar reflection. The
gas itself, the alarm quickly activates, often within milliseconds. UV detector is typically a photo discharge tube that detects deep
A second measure of protection is direct gas detection by UV radiation in the 180 nm to 260 nm wavelength range. Due to
means of catalytic-combustible gas detectors. They have a long atmospheric absorption, solar radiation at these wavelengths does
history and have been used for hydrogen applications for more not reach the earth’s surface; thus, the UV detector is essentially
than 50 years. The sensing devices have a pair of platinum-wire immune to solar radiation. The combination of IR and UV detec-
coils embedded in a ceramic bead. The active bead is coated with tion improves false alarm immunity, while producing detectors that
a catalyst, while the reference bead is encased in glass and is inert. can sense even small hydrogen fires at a 15-m range.
On exposure to hydrogen, the gas begins to burn at the heated Ultrasonic gas leak detection, catalytic gas detection, and
catalyst surface per the reaction: hydrogen flame detection have different strengths and vulner-
abilities. They respond to different hazard manifestations—the
2H2 ⫹ 2O2 r 2H2O ⫹ O2
The hydrogen oxidation releases heat
causing the wire’s electrical resistance to
change. The resistance is linear across a wide Hydrocarbon
temperature range (~500°C – 1,000°C) and Reactor 1 Reactor 2
mixture
proportional to concentration. For hydro-
gen-specific catalytic detection, the reac-
Stabilizer Light-
tion temperature and catalyst are tailored Hydrogen end
to prevent the combustion of hydrocarbons separator gas
in the substrate. The scheme’s simplicity Hydrogen mixture
makes catalytic detectors suitable for many
applications. Where gas accumulations may Gas detector
occur, catalytic sensors establish hydrogen Ultrasonic gas leak detector
presence with fair accuracy and repeatabil- Hydrogen and Reformate
ity. Hydrogen-specific catalytic detectors Flame detector hydrocarbon
mixture
also have fast response times (5s–10s) and
offer good selectivity. These parameters vary FIG. 4 Dual-stage reforming unit schematic that shows possible gas and flame detector
widely among the various manufacturers, locations.
46
gas, gas source or fire. Further, each technology operates in a dif- rising production must be matched by a comprehensive approach
ferent area of regard, with catalytic detectors as point instruments, to plant safety. New facilities that use hydrogen should be designed
and ultrasonic leak detectors and hydrogen flame detectors as area with adequate safeguards from potential hazards; the design of old
monitors. Due to their unique properties, combining detectors facilities should also be revisited to ensure that sufficient barriers are
increases the odds that hydrogen gas dispersal or fire is identified available to minimize accidents and control failure. Safety systems
early, either before ignition or when an explosion occurs. that deploy a diversity of detection technologies can counteract
An illustration using these technologies can be found in catalytic possible leak effects, fire and explosions, thus preventing equip-
reforming.1 In this process, a stream of heavy gasoils is subjected to ment or property damage, personal injury and loss of life.
high temperature (480°C–524°C) and pressure (1,379 kPa–3,447 A combination of catalytic and ultrasonic gas leak monitors
kPa; 200 psi–500 psi) and passed through a fixed-bed catalyst. and fire detectors is particularly effective because they are com-
Upon reaction, the oils are converted to aromatics that yield much plementary. The vulnerabilities of one are offset by the other’s
higher octane ratings for gasoline. Due to operating conditions and strengths, so there is less chance of propagating undetected haz-
the continuous production of hydrogen, a rupture in the reactors, ards. Such diverse safety systems, combined with a design that
separator or unit pipe system can have grave consequences. A detec- prevents leakage and eliminates possible ignition sources, offer a
tor allocation across a reforming unit is shown in Fig. 4. sound approach for managing hydrogen processes. HP
The scheme shown in Fig. 4 does not preclude the use of other
LITERATURE CITED
detection systems. Nor does it eliminate the need for operating 1 Berger, W. D. and K. E. Anderson, Modern Petroleum: A Basic Primer of the
procedures, instrumentation and control systems, and adequate Industry, Second Edition, PennWell Publishing, Oklahoma, 1981.
training—all necessary for safety. Condition monitoring instru-
ments, like X-ray pipe-testing equipment, play a pivotal role in
spotting defects before the pipe network integrity is lost. Likewise, Edward Naranjo is a product manager for General Monitors,
thermal conductivity sensors can ensure detection coverage under Inc. He has been with GMI for four years and contributes to product
innovation and new product development, including gas imaging
oxygen-deficient environments and thus complement catalytic sen- and ultrasonic technology initiatives. Mr. Naranjo has over 12 years
sors when used above the lower explosive limit. Experience suggests of product development experience in the industrial instrumenta-
the choice of detection instruments must be carefully weighed to tion, healthcare and consumer packaged goods industries. He received a BS degree
match the types of hazards associated with chemical processes at in chemical engineering from the California Institute of Technology and a PhD in the
same discipline from the University of California, Santa Barbara. Mr. Naranjo also
the refinery, and that each offset the other’s vulnerabilities. earned an MBA from the University of Chicago. He is the past chapter president
Hydrogen production will continue to grow, fueled by environ- of the Southern California Chapter of the Product Development and Management
mental legislation and demand for cleaner, higher fuel grades. But Association and is a certified new product development professional.

47
BONUSREPORT GAS PROCESSING DEVELOPMENTS
Fine-tuning demercaptanization
process: A case study
Optimizing caustic concentrations and reactor temperatures improved
acidic compound removal without installing new equipment
Z. MALLAKI , Sharif University of Technology and Petro Pars Ltd., Tehran, Iran; and
F. FARHADI, Sharif University of Technology, Tehran, Iran
L
iquefied petroleum gas (LPG) is often contaminated with Methods and materials. The propane and butane treatment
acidic compounds such as hydrogen sulfide (H2S), carbon and drying units of SPGC Phases 4 and 5 are designed to process
dioxide (CO2), carbonyl sulfide (COS), carbon disulfide sour propane and butane in two parallel identical trains; each train
(CS2), and methyl and ethyl mercaptans (thiols). Mercaptans in processes 50% of the total feed. Design capacity for each train is
lighter feeds, such as C3s, C4s, LPG and naphtha, are extracted 26,350 kg/hr and 41,100 kg/hr of sour butane and sour propane
with caustic solution processes, which are also referred to as cuts, respectively. The unit is designed to handle 40% to 100%
“sweetening” processes. of its normal capacity.1–3
Sweetening processes are widely applied to remove acid com-
pounds before transporting LPG for sale purposes. Tighter envi- S content and specifications for main equipment.
ronmental rules now require reducing the sulfur content of LPG Propane feed contains methyl mercaptan and COS with small
to 30 ppm. In this case study, an investigation is conducted to find amounts of ethyl mercaptan and only traces of H2S (less than
a cost-effective method to treat 1.2 wt% (12,000 ppm) sulfur- 1 ppm). The butane feed contains ethyl mercaptan, with small
content LPG streams to less than 30 ppm. amounts of methyl mercaptan and only traces of H2S and COS
(less than 1 ppm). Table 1 summarizes the design values for S
Background. LPG sweetening is a widely applied process using content of the feed. Table 2 lists the current mercaptans content
caustic to remove acid compounds from hydrocarbon streams. LPG as measured in January 2007. There is a significant decrease in the
desulfurization units of Iran’s SPGC Phases 4 and 5 were designed feed mercaptan amount as compared to the design specifications.
and constructed to produce sweetened propane and butane with a According to Table 2 and other plant data collected between
sulfur (S) content of less than 80 ppm via caustic extraction. Due 2006 and 2007, the total maximum amounts of mercaptans in
to stricter environmental regulations, these units could not meet butane and propane feed are approximately 2,300 ppmw and
new 30-ppm S content levels without modification. Although LPG 300 ppmw respectively. Specifications of the sweetening unit are
demercaptanization by caustic is widely applied in refineries, the listed in Table 3.
basic information necessary to optimize LPG units was missing. This
study was initiated to identify important performance factors for the Mercaptan extraction. When hydrocarbon and caustic
two existing sweetening/desulfurization units. Thus, optimization phase are intimately contacted, the mercaptans are absorbed into
requirements and consequent benefits were considered.
TABLE 2. Current amount of mercaptans in sour
propane and butane, SPGC Phases 4 and 5
TABLE 1. Sulfur impurities of sour propane and butane
Total mercaptans Total mercaptans
for design case, SPGC Phases 4 and 53 in sour propane in sour butane
Feed Sour propane Sour butane Date Train 1 Train 2 Train 1 Train 2
H2S, ppmw normal Trace Trace (<1 ppm) 1/1/2007 223.5 183.5 1,397
design Trace Trace 2/1/2007 2,306
COS, ppmw normal 167 Trace (<1 ppm) 3/1/2007 295.8 201.5 1,820
design 118 Trace
4/1/2007 1,824
Sulfur C3SH, ppmw normal 645 1,258
5/1/2007 220.6 163.5
impurities design 957 840
6/1/2007 1,522
C2H5SH, ppmw normal 59 11,300
design 31 8,000 7/1/2007 1,812
C3+ normal Trace Trace 8/1/2007 266.7 205.6
mercaptans, ppmw design Trace Trace 10/1/2007 2,090
48
GAS PROCESSING DEVELOPMENTS BONUSREPORT
TABLE 4. Constants A and B for Eqs. 1 and 2 4

Caustic Sand To caustic C4 Mercaptan structure A B
Extractor settler filter wash column
Caustic Caustic Caustic
molarity: 4.25 molarity: 2.97 molarity: 1.85
Mole Methyl mercaptan, 0.20235 33.7160 33.6521 33.5074
sieve CH3SH
dryers
Ethyl mercaptan, 0.05715 33.0043 32.9154 32.7771
LPG feed C2H5SH
Dry, sweet LPG
CW Propyl mercaptan, 0.02398 32.135 32.117 32.020
C3H7SH
To oxidizer
Caustic from Butyl mercaptan, 0.01617 31.297 31.28 31.263
wash column C4H9SH
Demineralized water
Where KE is extraction coefficient considering acid ionization
FIG. 1 Simplified process flow diagram of the extraction and is defined as:
section.11
(RS )aq + (RSH)aq
KE =
TABLE 3. Specifications of the main equipment (RSH)oil
Equipments Operating Operating Constants A and B are available in Table 4. Constant B in Table
temperature, °C pressure, barg
4 depends not only on mercaptan structure but also on caustic
Extraction No. of equilibrium stages molarity. Using experimental, constant B is developed by Eqs. 3
Propane extractor 40 29.5–31.5 15 and 4 for C1 and C2 mercaptans:4
Propane post- 70 30 7
treatment column B = 0.3504Ln(M ) + 33.267 for methyl mercaptan (3)
Butane extractor 40 11.1–13.3 15
B = 0.3112Ln(M ) + 32.571 for ethyl mercaptan (4)
Regeneration Dimensions: DⴛL (m2)
Oxidizer 50 5.5–6.0 1.4⫻14.3
As sodium mercaptides form in the caustic solution, the solu-
tion’s ability to extract mercaptans decreases, due to salting out.
DSO separator 50 5.6 2⫻10
The salting out effect is best represented by Eq. 5:4
C4 washing drum 40 15.3 1.6⫻5
S
log o = KC (5)
Sc
the caustic solution—sodium hydroxide (NaOH). Mercaptan
distribution between two phases—water and hydrocarbon— So Solubility in water
occurs as: Sc Solubility in salt solution
I II C Salt concentration in water
RSH RSH RS– K Salting-out constant
(Oil phase) (Aqueous phase) (Aqueous phase)
K = 0.075 For ethyl mercaptan
After extraction of mercaptans by the caustic solution, sodium K = 0.181 For n-butyl mercaptan
mercaptides are formed via this reaction equation:
Caustic regeneration. Rich caustic solution, leaving the
RSH + NaOH RSNa + H 2 O
extractor, is directed to an oxidizer, and air is injected into this
Fig. 1 is a simplified process flow diagram of the extraction sec- stream. The mixture flows upward through the oxidizer where
tion. According to experimental data represented for normal butyl alkaline is regenerated by conversion of sodium mercaptides to
mercaptans and assuming that variation of Kp and KE of C1 to C3 disulfides with CoSPc (sulfonated cobalt phthalocyanine) as
mercaptans with caustic molarity as well as temperature is similar catalyst. The separated alkaline solution is recirculated to the
to that of butyl mercaptan, empirical Eqs. 1 and 2 are represented extractors. In this process, the catalyst and alkaline solution are
for KE and Kp of C1 to C4 mercaptans for two liquid phases of regenerated (Eq. 7) and recycled:
isooctane and caustic solution. 4 These equations have shown
2RSNa + 0.5O2 + H 2O RSSR + 2NaOH
good agreement with experimental data of C1 to C3 mercaptans
extraction via caustic: Fig. 2 is simplified process flow diagram for the caustic regenera-
log K p = 5.856103 logT + A (1) tion section. Using experimental data represented in the literature,
the kinetic equation of mercaptide oxidation in an alkaline medium
Where Kp is the partition coefficient and is defined by Eq. 2 by molecular oxygen is developed as a function of temperature:5–7
when the pH is low enough to prevent acid ionization:
K 1K p [RS][Kt ][O 2 ]
[RSH]aq RSNa = 2.7667106
Kp = Since [RS ] = 0 1+ K p [O2 ]+ K r [RSSR]
[RSH]oil
log K E = 12.305 logT + B (2) exp(0.0385T ) (6)

I 49
Sour LPG to Air purge Temp: 40°C

gas plant Caustic to propane ratio: 0.1158 - Caustic mass flowrate: 4,761 kg/hr
Caustic to butane ratio: 0.2061 - Caustic mass flowrate: 5,358 kg/hr
Mercaptan remaining in propane or butane, ppmw

1,000
Disulfides
to storage
100
Lean caustic
to extractor
LP
steam 10
From extractor Oxidizer
Air
1
Fresh cat.
inject. syst. Spent caustic
to sump-drum 0.1
From process
air compressor
0.01
Sweet LPG cut 11 12 13 14 15 16 17 18 19 20 21 22
Caustic concentration, wt x 100
FIG. 2 Advanced process flow diagram of extraction section.11 Ethyl mercaptan into the butane extractor: 12,800 ppmw
Ethyl mercaptan into the butane extractor: 3,500 ppmw
Methyl mercaptan into the propane extractor: 690 ppmw
TABLE 5. Minimum required caustic concentration Methyl mercaptan into the propane extractor: 330 ppmw
for product purity under specified conditions
Ethyl mercaptan in the Minimum required caustic FIG. 3 Simulation results of propane and butane purity vs. caustic
butane product, ppmw concentration, wt% concentration in sweetening process.
4.5 16
5 15.5 increase greatly due to the caustic molarity. Experiments showed
5.5 14.93 that up to caustic concentration of 2.75 molar mercaptans conver-
10 13.6 sion to mercaptides will rapidly reach to 92%; thereafter, increasing
20 12.4
the caustic concentration is not so important.9
Simulation results of propane and butane purity vs. caustic
30 11.8
concentration are presented in Fig. 3 for design and actual operat-
80 10.2 ing conditions. For caustic concentrations greater than 13 wt%,
Ethyl mercaptan in sour butane, ppmw 2,500 the mercaptan content of the propane products was reduced below
Temperature, °C 40 0.5 ppm. Table 5 summarizes the minimum required caustic con-
Mass ratio of caustic solution to butane 0.2061 centration to reach specific product purity for assumed mercaptan
Caustic flowrate, kg/hr 26,340.1 content and conditions. To process present mercaptan content for
sour propane and butane, the optimum caustic concentration is
14.93 molar. Thus, the mercaptan impurity will fall to 0.1 ppmw
The constants in Eq. 6 are: and 5 ppmw in propane and butane products, respectively.
K1Kp = 2.07 ⫻ 10–2 m3 / [Pa-mole-s] However, for the normal design case in Table 1, Fig. 3 shows
Kp = 1.1 ⫻ 10–4 Pa–1 that, by applying a caustic concentration of 14.93 wt%, under
Kr = 950 m3/mole specified conditions, only 0.3 ppmw and 50 ppmw methyl mer-
The concentration of mercaptide ion [RS–], catalyst [Kt] and captan and ethyl mercaptan remains in the treated propane and
disulfides [RSSR] are expressed in mole/m3. The concentration of butane products, respectively.
oxygen [O2] is specified in Pa. The term [RSSR] reflects the inter-
facial mass transfer effects and is determined from the experimen- Temperature and extraction efficiency. Results from
tal data of [RSH]oil vs. time and then calculated by subtracting the experiments treating butyl mercaptan with two liquid phase of 0.5
[RSH]oil at a particular time from the initial concentration.6–8 molar caustic and isooctane at different temperatures shows that the
partition coefficient (Kp) is independent of temperature and mercap-
Molecular structure. According to the experimental data, tan ionization constant decreases with lower temperatures. However,
although increasing the molecular weight of mercaptans has neg- the extraction coefficient is enhanced with decreasing temperatures
ligible influence on ionization constant, it decreases mercaptan since the hydrolysis (Kh=Kw /KA) constant likewise decreases:4
solubility in water and thus KE. [H+ ][OH ]
Kw = = [H+ ][OH ] Water ionization
[H 2 O] constant
Caustic concentration and extraction efficiency.
Increasing caustic molarity will increase the extraction coefficient. [RS ][H+ ]
However for C3+ mercaptans, this effect increases up to a caustic KA = Mercaptan ionization
[RSH] constant
molarity of 3. After this point, the salting out phenomena occurs.
Kh = Kw / K A Hydrolysis constant
Thus, the partition coefficient (Kp) decreases, and the KE does not
50
TABLE 7. Minimum practical caustic flowrate

Caustic concentration: 14.93 wt% according to product purity, under specified conditions
Caustic to propane ratio: 0.1158 - caustic mass flowrate: 4,761 kg/hr
Caustic to butane ratio: 0.2061 - caustic mass flowrate: 5,358 kg/hr Methyl mercaptan in the Mass ratio of caustic
10,000
propane product, ppmw solution to propane
Mercaptan in propane or butane product, ppmw
1,000 0.1 0.1158

1 0.1020
100 5 0.0930
10 0.0890
10
30 0.0800
1 Ethyl mercaptan in sour butane, ppmw 330

Temperature, °C 40
0.1 Caustic concentration, wt% 14.93
Caustic flowrate, kg/hr 41,113.22
0.01
TABLE 8. Minimum practical caustic flowrate

0.001
according to product purity under specified conditions
0.0001
Ethyl mercaptan in the Mass ratio of caustic
15 20 25 30 35 40 45 50 55
Temperature, °C butane product, ppmw solution to butane
5 0.2060
Ethyl mercaptan into the butane extractor: 3,500 ppmw 10 0.1960
Ethyl mercaptan into the butane extractor: 2,400 ppmw 20 0.1804
Methyl mercaptan into the propane extractor: 690 ppmw
Methyl mercaptan into the propane extractor: 330 ppmw 30 0.1708
Ethyl mercaptan in sour butane, ppmw 2,500
FIG. 4 Purity of the propane and butane products as a function of Temperature, °C 40
temperature based on simulation results. Caustic concentration, wt% 14.93
Caustic flowrate, kg/hr 26,340.1
TABLE 6. Maximum practical temperature according
to product purity under specified conditions
Ethyl mercaptan in Maximum practical will precipitate out of the caustic solution and possibly cause
butane product, ppmw temperature, °C plugging problems.
5 40 The upper temperature limit is 45°C, because the mercaptan
10 41.5
extraction efficiency begins decreasing. Since temperatures of the
sour propane and butane from the NGL fractionation unit are 60°C
30 43
and 40°C, respectively, the optimum temperature of 40°C for both
80 45 extractors is recommended to achieve less than 10 ppmw mercaptan
Ethyl mercaptan in sour butane, ppmw 2,500 concentration in the product under specified conditions.
Caustic concentration, wt% 14.93
Mass ratio of caustic solution to butane 0.2061 Caustic flowrate and extraction efficiency. Caustic con-
Caustic flowrate, kg/hr 26,340.1 sumption—kg of 100% NaOH per metric ton of feedstock—for a
given treating level is directly related to the initial caustic solution
concentration, initial mercaptan concentration in the feedstock and
According to the experiments, mercaptan extraction is favored at product purity. Experiments with refinery tests on LPG demercapta-
lower temperatures. Simulation results of propane and butane purities nization units have confirmed that, if the mercaptan content entering
vs. temperature are presented in Fig. 4 for design and actual operating an equilibrium stage, is very high, then NaOH solution saturation
conditions. As expected, reducing process temperature will improve with mercaptans is a limiting factor for extraction.10 Studies have
mercaptan extraction. Temperatures lower than 44°C yield a mer- shown that the saturation value, expressed in moles S–2 per mole
captan content of less than 1 ppmw in the propane product, under NaOH, does not depend on the initial mercaptan content in the
the specified conditions in Fig. 4. Maximum practical temperatures product being treated. This saturation value decreases with increas-
for butane products with different specifications are summarized in ing initial caustic solution concentration. For a given treating level
Table 6. and NaOH solution concentration, the saturation value is constant.
Data from Table 6 illustrate the significance of temperature Considering the saturation capability of caustic solution as
control in the butane extractor. Although reducing temperature a function of caustic concentration, Eqs. 7 and 8 are regression
will enhance extraction efficiency; other processing effects are equations that describe the experimental data:10
possible:
Y 2 = 0.350 0.00X 1 (7)
• For temperatures lower than 20°C, caustic entrainment
problems will occur. Y2 Saturation of the caustic solution for averaging, moles
• At lower temperatures, sodium sulfide and carbonate salts S –2/mole NaOH
I 51
Temperature: 40°C - caustic concentration: 14.93 wt% -

Temperature: 40°C - caustic concentration: 14.93 wt% - butane mass flowrate: 26,340.1 kg/hr
propane mass flowrate: 41,113.22 kg/hr 1,000
1,000
Ethyl mercaptan remaining in butane product, ppmw

Mercaptan remaining in propane or butane product, ppmw
100
100
10
1
10
0.1
1
0.01
0.15 0.17 0.19 0.21 0.23
0.07 0.08 0.09 0.10 0.11 0.12
Mass ratio of caustic solution to butane
Mass ratio of caustic solution to propane
Methyl mercaptan into the propane extractor: 690 ppmw Ethyl mercaptan into the butane extractor: 3,500 ppmw
Methyl mercaptan into the propane extractor: 330 ppmw Ethyl mercaptan into the butane extractor: 2,500 ppmw
FIG. 5 Purity of the propane and butane products as a function of caustic flowrate based on simulation results.
X1 NaOH weight fraction in caustic solution ⫻ 100 ■ Using operating data, engineers ran
Y 2 = 0.624 0.016X 1 (8) simulation models that more accurately
Y´2 Saturation of the caustic solution for breakthrough, respresented the ‘sweetening’ process for
moles S–2/mole NaOH
X1 NaOH weight fraction in caustic solution ⫻ 100 this gas plant.
Simulation results shown in Fig. 5 and Tables 7 and 8, represent The colorimetry of the CoSPc—a reliable means for deacti-
the required caustic (NaOH) amount based on the impurities lev- vation measurement—shows that the catalyst activity at room
els before and after treatment, under the specified conditions. temperature is greater than that of higher temperatures. From the
Based on these results, 0.102 kg of caustic solution of 14.93 literature, adding catalyst to previously prepared caustic solution
wt% (0.015 kg pure NaOH) per kg of propane and 0.210 kg of can provide the highest conversions.8
caustic solution of 14.93 wt% (0.032 kg pure NaOH) per kg of
butane guarantee propane product and butane product with mer- Air injection and caustic regeneration efficiency.
captan impurities of 1 ppmw and 5 ppmw, respectively. Result: The stoichiometric amount of oxygen to oxidize sodium mer-
Higher purity marketable products are now available. captides is 0.25 mole of oxygen per mole of sodium mercaptide.
However, it is necessary to inject excess air into the oxidizer to
Mercaptan structure and regeneration efficiency. enhance reaction efficiency. This excess air depends on the sodium
From experimental reaction results for several sodium mercaptides mercaptides concentration in the inlet caustic solution.
with different structures at similar conditions, it can be found For an initial mercaptide content of 35,770 ppm at the inlet,
that the more complex the structure of sodium mercaptide the approximately 200% excess air is needed to reach to 5 ppmw
slower the oxidation rate.2 Tert-butylmercaptide is one of the ethyl mercaptide content at the outlet. Considering the actual
most difficult mercaptides to be oxidized due to its high steric conditions, 1.16% excess air will yield the same ethyl mercaptide
and inductive effects.8 concentration (5 ppmw) in the caustic solution, leaving the reac-
tor if 8,680 ppm of mercaptide is associated with the feed entering
Stability in LPG sweetening. With continuous unit opera- the reactor. However, there are some key points:3
tions, the catalyst will deplete; sweetening efficiency will deterio- 1) While a low mercaptan concentration is desirable, the
rate and the alkaline solution must be replaced frequently. This caustic solutions should never be completely regenerated via high
will increase operating costs as well as cost for waste disposal of excessive air rates. In the absence of mercaptans, traces of oxygen
the alkaline solution. can dissolve in the circulating caustic and cause sweetening to
52
1,000 40
Ethyl mercaptide remaining in the

39
regenerated caustic, ppmw

100 38
Ethyl mercaptide remaining in
37
regenerated caustic, ppmw
36
10 35
34
33
1
32
31
0.1 30
29
28
0.01 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18
Mass fraction of NaOH in the caustic solution, xw
0.001
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 Inlet caustic containing 8,234 ppm ethyl mercaptide:
Excess air, mole air injected/mole air stochiometric Oxidizer temperature from top to bottom:40°C-50°C
Excess air: 110%
Ethyl mercaptide into the oxidizer: 8,234 ppmw Mass flowrate of caustic solution: 13,851.9 kg/hr
Ethyl mercaptide into the oxidizer: 35,770 ppmw
FIG. 7 Sodium mercaptide in regenerated caustic vs. caustic
For caustic solution containing 8,234 ppm ethyl mercaptide: concentration based on the simulation results.
Required Inlet concentration of caustic:12.5 wt% (dealing with not
prewashed propane)
Outlet concentration of caustic: 14.93 wt%, oxidizer top to bottom TABLE 9. Variations of caustic molarity by mass
temperature:40°C-50°C
fraction of sodium hydroxide in solution
Mass flowrate of caustic solution:13,851.9 kg/hr
For caustic solution containing 35,770 ppm ethyl mercaptide:
Required Inlet concentration of caustic:14.54 wt% Caustic solution mixed with 110%
(if propane is prewashed) Pure caustic solution excess air under conditions of Fig. 7
Outlet concentration of caustic: 14.93 wt%, oxidizer top to bottom Mass fraction Molarity of NaOH Mass fraction Molarity of NaOH
temperature: 40°C-50°C of NaOHⴛ100 in the solution of NaOHⴛ100 in the solution
Mass flowrate of caustic solution: 14,300.3 kg/hr
17.8 5.223 17.1 2.18
FIG. 6 Sodium mercaptide in regenerated caustic as a function of 14.9 4.288 14.5 1.98
excess air based from simulation results. 14.1 4.016 – –
12.9 3.626 – –
occur in the extractor; the disulfides will then return to the LPG 11 3.007 11.3 1.78
phase and increase the product’s total sulfur content.
2) Small levels of mercaptides in the caustic (30 ppmw–50 ppmw)
keep the catalyst dispersed. Thus, the catalyst does not accumulate at caustic molarity by mass fraction of NaOH in solution are not
the rich-disulfide caustic interface in the disulfides separator. the same as molarity variations of pure caustic solution by its
Consequently, regenerated caustic must hold 30 ppmw–50 composition, as listed in Table 9.
ppmw sodium mercaptide. In Fig. 6, the present unit operates Since regenerated caustic is recycled from the oxidizer to the
with 8,234-ppmw sodium mercaptide concentration at the inlet extractors, the concentration of regenerated caustic at the reactor
of the oxidizer, and 108%–110% excess air is the optimum value. outlet must be the same as the caustic concentration entering the
The oxygen level in the air leaving DSO separator must range extractor. Caustic concentration at the reactor inlet is specified
between 1.5% and 2%. as a function of the sodium mercaptides concentration to be
oxidized to NaOH and the caustic concentration at the inlet of
Caustic concentration and regeneration efficiency. the extractors.
Caustic solution as a reaction medium has an optimum concen- For present plant conditions, 8,200 ppmw of sodium mercap-
tration of 1.8–1.9 molar, which supports 75% conversion. While tide is oxidized to NaOH. The optimum caustic concentration to
increasing the caustic concentration to 3.8 molar is still practical; the extractors and, thus, recycling from the oxidizer is 14.93%.
the high levels only yield 70% conversion. Consequently, very Consequently, the caustic concentration from the extractors to
high caustic concentrations are not beneficial to regenerating the oxidizer must be increased from 12.2 wt% to 14.5 wt% at the
NaOH. To explain the regeneration reaction kinetics, there are oxidizer inlet. Accordingly, 884.41 kg/hr of fresh caustic (solution
two points. First, when increasing the concentration of caustic of 40% wt) makeup is mixed with the rich-caustic solutions from
solution, the solubility of CoSPc catalyst will decrease catalyst the extractors. Referring to Fig. 7, 30 ppmw of sodium mercaptide
dispersion in the solution. Second, higher alkaline solutions have will remain in caustic solution, which is a desirable level. Note:
greater viscosities, which hinders the transfer of free radical in the SPGC Phases 4 and 5 propane is not prewashed; thus, a large
radical oxidation reaction of mercaptides.7,8 volume of fresh caustic is required.
Experimental results suggest an appropriate alkaline concen-
tration of 2.75–4.25mol/dm3 for the sweetening of LPG. Fig. 7 Temperature and efficiency of caustic regeneration.
shows the simulation results over the effect of caustic concentra- Temperature is one of the most important factors influencing
tion on the rate of mercaptide oxidation. Rich caustic solution reactions. To oxidize propane mercaptide, the optimum tempera-
in the oxidizer is mixed with air as oxidant. Thus, variations of ture based on oxidizer performance ranges between 40°C–50°C.
I 53
required concentration of recycling caustic to the extractor as

100 well as the amount of sodium mercaptide impurities in the rich
caustic are limiting factors for the unit and should be considered
when defining the required concentration of the inlet caustic to
Ethyl mercaptide in caustic
leaving the reactor, ppm
the oxidizer.
High-purity propane and butane products were obtained in
SPGC Phases 4 and 5 when operating variables were adjusted.
10
LPG with mercaptan content less than 10 ppm is sold at $ 3/
ton to $4/ton—more than present LPG prices. Consequently,
optimizing this unit resulted in a total net income increase of
$2.9–$3.9 million/yr. This task is achieved without new equip-
ment installed or equipment modifications. The results were
1 possible by only fine-tuning operational process parameters with
20 25 30 35 40 45 50 55 some extra caustic consumption reduction. HP
Mean log temperature, °C
Inlet caustic containing 8,234 ppm ethyl mercaptide: NOMENCLATURE
Required inlet concentration of custic: 12.5 wt (if propane is
not prewashed Kp Partition coefficient
Excess air: 110% KE Extraction coefficient
Mass flowrate of caustic solution: 13,851.9 kg/hr So Solubility in water
Inlet caustic containing 35,770 ppm ethyl mercaptide: Sc Solubility in salt solution
Inlet concentration of caustic: 14.54 wt% (if propane is C Salt concentration in water
prewashed) K Salting-out constant
Excess air: 200% rRSNa Reaction kinetic of sodium mercaptide oxidation
Mass flowrate of caustic solution: 14,300.3 kg/hr KW Water ionization constant
KA Mercaptan ionization constant
FIG. 8 Sodium mercaptide in regenerated caustic as a function of Kh Hydrolysis constant
temperature based on the simulation results. Y2 Saturation of the caustic solution for averaging
(moles S–2/mole NaOH)
Y´2 Saturation of the caustic solution for breakthrough
However, the oxidizer temperature should always be kept as low (mole S–2/mole NaOH)
as possible considering catalyst activity while still maintaining X1 NaOH weight fraction in caustic solution⫻100
the desired degree of mercaptans regeneration. In any event, T Temperature
55°C would be considered as an absolute maximum temperature xw Weight fraction
because of metallurgical limitations and also the possibility of
disulfide oils decomposing into sulfonic acids. ACKNOWLEDGMENT
Based on simulation results, Fig. 8 shows the effect of oxi- The authors thank South Pars Gas Company R&D for their support and their
dizer temperature on the conversion of sodium mercaptide. The permission to publish this article.
results are presented for two cases—design and actual operating
conditions. The extraction of 2,500 ppmw of ethyl mercaptan LITERATURE CITED
1 C. P. D., Propane Treatment, Operating Manual, Chapter 2, Process Section 2,
from butane and 330 ppmw of methyl mercaptans from pro-
Iran South Gas Field, Phases 4 and 5, Unit 114, June 2003.
pane by caustic will yield 8,234 ppmw of sodium mercaptide in 2 C. P. D., E. L., Butane Treatment, Operating Manual, Chapter 2, Process
the caustic solution at the oxidizer inlet (Fig 8). Since this is an Section 2, Iran South Gas Field, Phases 4 and 5, Unit 115, June 2003.
endothermic reaction, if the sodium mercaptide content of the 3 d’ESTEVE, C., “Sulfrex Process, Process Data Book, South Pars Phases 4 and
caustic at the reactor inlet is 8,234 ppmw, then the reactor top 5,” On Shore Facilities, Assaluyeh, p. 7, pp. 20–21, 2001.
4 Aminian, H., “Chemical refining of condensate produced by Iran’s Razi
and bottom optimum temperatures should be approximately Complex,” M Sc. Thesis, Sharif University of Technology, pp. 25–34, 1996.
45°C and 50°C, respectively under mentioned conditions in 5 Mazgarov, A., “Desulfurization of Oil, Gas, Petroleum Products and
Fig. 8. Remember: At least 30 ppm of RSNa must remain in Wastewater,” Volga Research Institute of Hydrocarbon Feed, Kazan, Russia,
regenerated caustic. 2005.
6 Mazgarov, A. M., “A selective treatment of various oils and gas condensates to
remove light mercaptans and hydrogen sulfide,” World Petroleum Congress,

Outlook. According to the results, caustic concentration of 2006.
14.93 wt% and temperatures of 40°C are optimum values for 7 Ruiting, L., X. Daohong and X. Yuzhi, “Oxidation of sodium mercaptide
extractors. The required amount of caustic to extract mercaptans with sulfonated cobalt phthalocyanine as catalyst,” American Chemical Society,
Vol. 48, No. 2, pp. 74–76, March 2003.
can be selected according to the purity of the product, as shown 8 Ruiting, L., X. Daohong and X. Yuzhi, “ Study on the Stability of CoSPc in
in Tables 7 and 8. When considering caustic regeneration con- LPG Sweetening,” American Chemical Society, Vol. 48, No. 4, pp. 338–340,
ditions, amount of air injection to the oxidizer is a key factor August 2003.
9 Ruiting, L., X. Daohong, X. Yuzhi and T. Yongliang, “Effects of caustic con-
affecting the sweetening process efficiency. Approximately 30
centration on the LPG sweetening,” Petroleum Science and Technology, Vol. 23,
ppm–50 ppm of sodium mercaptide must be included in the No. 5–6, pp. 71–72, May/June 2005.
circulating caustic. Fig. 6 shows the required air amount for 10 Tukov, G. V., N. N. Ivanova, A. N. Sadykov, A. M. Polotskii and N.
specified conditions. The optimum log mean temperature of A. Glebova, “Establishing Standards for Consumption of Caustic Soda
the oxidizer is 40°C to 45°C depending on the impurities con- in Treating Liquefied Petroleum Gases (LPG) to Remove Mercaptans,”
Chemistry and Technology of Fuels and Oils, Vol. 11, No. 11–12, pp 869–872,
centration. The optimum caustic concentration of 1.9 molar November/December 1975.
after mixing with air is the optimum value within the oxidizer, 11 Savary, L., “Gas Processing with Axens’ Technology, From Purification to
which can be adjusted by fresh caustic makeup. However, the Liquefaction,” Axens, 1996.
54
What are the opportunities

to construct liquefaction facilities
at the Arctic Circle?
Building and operating natural gas plants in the high latitudes pose
numerous challenges
D. A. WOOD and S. MOKHATAB, David Wood & Associates, Lincoln, UK
L
ocating natural gas liquefaction installations around the Arctic continental shelf are limited to a 200-nautical mile (nm)—approx-
Ocean for export markets poses many challenges. This region imately 370-km—economic zone bordering their coasts. Under
is hostile with many changing environmental obstacles. As international law, no country can claim sovereignty to the areas
shown in Fig. 1, many hurdles must be addressed when construct- surrounding the North Pole. The 1982 United Nations Conven-
ing and operating a liquefied natural gas (LNG) facility. Yet, the tion on the Law of the Sea (UNCLOS) provides a country with
potential oil and gas resources located at the Arctic region draw a 10-year period to make claims to extend its 200-nm zone. Due
global interest. Several formidable obstacles must be addressed in to this, Norway (ratified UNCLOS in 1996), Russia (ratified
conquering this region to develop these new energy resources. UNCLOS in 1997), Canada (ratified UNCLOS in 2003) and
Denmark (ratified UNCLOS in 2004) have launched claims
Arctic Ocean and its margins. The Arctic Ocean is a vast, under the convention that certain Arctic sectors should belong
remote and inhospitable region. A substantial portion of its con- to their territories.1 The US has signed, but not yet ratified this
tinental shelf lies off the north coast of Russia, which is where treaty. Because of the potential mineral resources possibly existing
most of the human settlements proximate to the Arctic Ocean
are located (Fig. 2). The North Pole is surrounded by the Arctic
Ocean. Five countries surround the Arctic Ocean: Russia, the US Pacific
(via Alaska), Canada, Norway and Denmark (via Greenland). Ocean
Currently, these nations’ claims to sovereignty over the Arctic Bearing

Okhotsk
Sea
Sea Anadyr
Anchorage
Whitehorse Fairbanks
Arctic challenges for the LNG industry to overcome Alaska
Pevek
(United States)
Sufficient Developing High-cost Transportation
Yellowknife
Safe operations yet-to-find sub-giant technologies through
Canada Holman
Tiksi
in extreme gas reserves field sizes variable

volumes sea ice Norilsk
conditions Talnah
Arctic Kajerkan Russia
Political Wide-ranging
seasonal Resolute Ocean
posturing Hudson Dikson Dudinkha
temperatures Bay Thulé Apatity
and Kirovsk Novy Urengoï
Environmental Rapidly Ivujivik Baffin Labytnangi
Nadym
footprint Bay Greenland Svalbard Kandalaksha Vorkuta
changing Iqaluit (Norway) Salekhard
(Denmark) Naryan
Project weather Illulissat Monchegorsk Mar Pechora
Legal investment Nuuk Kangerlussuaq Murmansk and
Indiga
disputes Modular Severomorsk
decisions multi-site
Tromsø
Norwegian Archangelsk
Sea Bodø and Novodvinsk
Regulatory parallel Rovaniemi
Severodvinsk
Reykjavik
framework engineering Iceland
Kiruna Finland Onega
Atlantic
Complex Ocean Feroe Islands Norway Sweden
Fiscal (Denmark)
upstream
terms
interfaces
Attracting skilled Intermittent Commercial Lower Fluctuating
Population in agglomerations
human resources delivery sustainability operating plant
schedules at low efficiencies operating 400,000
100,000 200,000 NB: The small blue dots represent villages with less
gas prices conditions 20,000 50,000 than 20,000 inhabitants and very small communities.
FIG. 1 Challenges of exploiting Arctic Ocean natural gas resources FIG. 2 The Arctic Ocean and its surrounding settlements. Source:
with LNG supply chains. UNEP/GRID-Arendal Maps and Graphics Library, 2005.9

I 55
World Arctic cumulative discovery 2010 – 2030 2040 – 2060 2070 – 2090
180 450
160 O+C Gb 400
Cumulative mean discovery, Gboe
G Tcf/6
Cumulative number of fields

140 Field 350
120 300
100 250
Ultimates
80 Oil 50 Gb 200
Gas 150 Gboe FIG. 5 Forecast impacts of warming Arctic: Arctic Climate Impact
60 = 1,000 Tcf
150
Assessment. Source: Cambridge, UK: Cambridge University
40 100 Press.11
20 50
0 0 shows extrapolated discovery trends of the second report, which
1940 1950 1960 1970 1980 1990 2000 2010 used mathematical models to estimate ultimate recoverable petro-
Cumulative number of new field wildcats leum reserves of 50 billion barrels of oil and 1,000 trillion cubic
feet (Tcf ) of natural gas for a combined 217 billion boe. Although
FIG. 3 World Arctic cumulative discovery of oil and gas resources
through to the end of 2006.3 this study excludes Greenland, it does highlight that most land
sections of the Arctic are already well explored and can be used
reliably to estimate yet-to-find resources.
+2 With much exploration to be undertaken, it is no surprise that
Observed temperatures yet-to-find estimates vary widely. However, there is a consensus
Temperature anomaly, °C
10-year running mean

+1 among analysts that approximately three-quarters of the reserves
in the Arctic Ocean sedimentary basins are natural gas. The major
0 oil and gas companies are attracted by the potential of finding
other giant fields such as the Shtokman in the Barents Sea.
-1 For the global gas consumers and long-term sustainability
of natural gas as a major global energy source, a more signifi-
-2 cant challenge is advanced technologies that can cost-effectively
1880 1900 1920 1940 1960 1980 2000 develop the numerous smaller-sized gas fields of the Arctic Region.
FIG. 4 Trends in Arctic temperature, 1880–2006. Source: These methods, in addition, could be applied to the few giant gas
CRUTEM3v dataset, Climate Research Unit, University of fields that remain undiscovered and could be developed using
East Anglia.10 existing technologies and resource approaches. In this case, the
technological focus should be on how to commercially develop and
in the deeper waters of this region and the ability to control stra- transport a large portion of these gas resources to global markets,
tegic shipping routes, there is significant competition and political not just on how to develop a few giant fields.
maneuvering by these nations to optimize the size of their claims. It
is therefore unlikely that clearly defined and internationally agreed Changing Arctic climate opens new frontier. Although
borders covering the entire Arctic Ocean region will be available in some debate remains over the causes of higher global temperatures,
the near future. Some resource development could be delayed due the evidence and consequences of climate change are nowhere
to potential international disputes over such borders. more evident than in the Arctic Ocean and its margins. The con-
sequences of a rising Arctic temperature trend (Fig. 4) according to
How much petroleum exists in the Arctic? There is scientific models are likely to be quite rapid and cause substantial
much uncertainty concerning the volumes of oil and gas that exist contraction of sea ice (Fig. 5). The continental margins of the Arc-
and can be commercially recovered from Arctic regions. Some tic Ocean are also likely to see environmental changes due to higher
speculate that between one quarter and one third of all remain- mean annual temperatures before the end of the century (Fig. 6).
ing oil and gas reserves to be found worldwide could possibly Contemporary conditions around the Arctic Ocean continen-
be located in the Arctic regions. A study by Wood Mackenzie tal shelf vary substantially. For instance, whereas the Barents Sea
reported a more conservative view that 233 billion barrels of oil remains ice-free even in winter (due to the influence from the Gulf
equivalent (boe) of oil and natural gas combined has already been Stream), the Chukchi Sea is ice-locked in winter. Changing marine
discovered in Arctic basins.2 It is estimated that some 166 billion currents could have significant consequences for local ice condi-
boe remain undiscovered (yet-to-find). That report identified the tions, and these are more difficult to predict. Accordingly, there
South Kara-Yamal basin and the East Barents Sea in Russia, along is much uncertainty over which regions will become navigable in
with Greenland’s Kronprins Christian basin to have yet-to-find winter by shipping, including LNG carriers. The Arctic Ocean
resources greater than 10 billion boe. However, only the South will, under all climatic scenarios, remain a challenging nautical
Kara-Yamal basin and the East Barents Sea were considered to environment to navigate and this will require special ship designs.
offer yet-to-find potential in pool sizes of over 1 billion boe. In terms of oil and gas operations, extreme cold and limited winter
An even more conservative view is expressed by the IHS data- daylight pose both operational and human endurance challenges.
base (February 2007) for existing Arctic fields and New Field The longer-term global consequences of such dramatic changes
Wildcats (NFW) for Russia, Europe (Norway and Svalbard) and in the Arctic Ocean (e.g., rising sea levels and less predictable weather
North America (US and Canada) north of 66°33’39’’.3 Fig. 3 patterns) are more difficult to forecast and may have significant overall
56
Projected perma
frost b
o un
da
ry
Observed sea-ice
September 2002
e
FIG. 7 Arctic LNG shuttle Höegh LNG. The photo is used with
lin
e
lin
permission from Höegh LNG.

tre
ree
Projected
ted
tt
en
ec
sea-ice
es
j
Pro
Pr
2070-2090 that are largely the challenges for LNG facilities and operating
equipment. Winterization technologies are required to restrict icing
ary
und
at the air and gas inlets and initial chilling plants, but these units
bo
can require frequent adjustments as weather conditions vary widely
t
os
a fr
leading to inefficiencies.5 The propane refrigerant cycle provides the
m
p er
nt initial chilling in the most commonly licensed liquefaction processes
rre
Cu
and is responsible for taking temperatures down to the –35°C to
–40°C. The cycle is also used to liquefy and separate substantial
volumes of gas liquids from the feed gas. To improve initial cooling
cycle efficiencies under Arctic conditions may require replacing
FIG. 6 Impacts of a warming Arctic. Source: Arctic Climate Impact propane as a refrigerant with a lower boiling point gas (e.g., ethane
Assessment (ACIA), 2004, and UNEP/GRID-Arendal Maps or ethylene) or a multi-component mixed refrigerant.
and Graphics Library.9,12 The ability of liquefaction plants to benefit from theoretical
higher efficiencies at cold temperatures depends upon the design
negative sustainability consequences. However, the medium-term temperatures for these Arctic plants and their design operating
implications of such scenarios are: a greater number of Arctic sea ports strategies. If the average annual temperature is used as a fixed-
will be ice free during the winter; a greater area of the Arctic Ocean design temperature, losses due to higher than average tempera-
will be navigable for shipping; and easier access to oil and gas resources tures (assuming a rate of 1.8%/°C) significantly outweigh gains
beneath the Arctic continental shelf. It is likely that countries and cor- attributable to more efficient condenser performance at lower
porations will make efforts to exploit such opportunities. than average temperatures, as plant capacities are varied to achieve
The potential of access to additional petroleum resources and annual production quotas.6 Conversely, fixing the design through-
the opening of a new exploration and development frontier are put capacity and raising design temperatures (above average ambi-
stimulating many in the petroleum industry. The energy industry is ent conditions) to achieve that capacity can lead to higher total
becoming excited about these opportunities and is seriously consid- efficiency, but at higher capital costs.5
ering the technological challenges associated with exploiting Arctic If liquefaction plants are to be operated at varying throughput
resources. One of the first indications of institutional cooperation capacities dependent on changing ambient temperatures, then the
is the agreement reached in April 2008 between the American feed gas and LNG shipping logistics must be adjusted to cope with
Bureau of Shipping and the Russian Maritime Register of Ship- such variations. This may not always be possible. For instance, colder
ping to jointly develop classification rules for Arctic LNG carriers.4 weather conditions may lead to shipping delays at a time when the
This agreement came in the wake of the Shtokman Development plant is capable of maximum output. The liquefaction plant opera-
Co. preparing plans for the giant Shtokman gas field (>100 Tcf of tors will have to balance the economic benefits of larger-capacity train
reserves) in the Barents Sea. Russia, following Norway’s SnØhvit installations, optimum design configuration from an operating per-
LNG project (onstream September 2007), is known to be plan- spective, and the challenges of constructing and operating the plant
ning substantial gas liquefaction facilities along its northern coast at remote sites under adverse and variable weather conditions.
to enable worldwide exports of its gas resources. Among the Rus- Limited winter daylight hours, more costly human resources and
sian oil industry’s plans under consideration is an LNG plant in difficult construction logistics also have to be acknowledged as major
Teriberka on the Barents Sea coast, along with a plant in the Yamal contributions to greater capital and operating costs and extended
Peninsula. Russian state-owned gas monopoly Gazprom and its project schedules. The very large cost overruns vs. the originally sanc-
subsidiary Sevmorneftegaz expect that 25 new LNG tankers will tioned budgets experienced by the StatoilHydro-operated SnØhvit
be required in connection with the Shtokman project. No surprise LNG plant, and the Shell-operated Sakhalin LNG plant during their
that the LNG shipping industry is showing interest. construction phases, and the significant and costly teething problems
experienced by the former testify that installing liquefaction plants at
Liquefaction at high latitudes. Cold average annual tem- high latitudes has substantial associated cost penalties.
peratures are actually beneficial for operating efficiencies and energy Modular and offsite construction of major components offer a
consumption by cryogenic facilities, regardless of the technology partial solution to some of these problems. But careful upfront plan-
applied. For example, cold ambient temperatures enhance gas- ning, extensive front-end engineering and design evaluations and
turbine power outputs. In the Arctic region, it is not, therefore, the parallel engineering, procurement and construction methodologies
average annual temperature, which is low (close to 0°C, the point at would be necessary to effectively execute such projects. Multi-site
which fresh water freezes), that poses the challenge to gas liquefac- operations themselves pose challenges due to resource procurement,
tion. Rather, it isthe seasonal temperature and weather variations integrated planning, control, regulatory and fiscal complexity.
I 57
Operations and maintenance issues. Winterization of than for ice-free supply chains. Shuttle-tanker methodologies may
gas processing and liquefaction plants is necessary to prevent fluid make sense in some cases, i.e., ice-classed tankers to move cargoes
freezing, liquid drop-out, and wax and hydrate formation. Elements past the ice edge either to trans-shipment ports or for ship-to-ship
of gas-processing plants, pre-cooling refrigeration cycles and air- transfer may make commercial sense in some cases. The reality is
cooling systems are most likely to experience such problems. Sys- that each port and shipping route will probably pose its own chal-
tems that facilitate rapid responses to short-term changes in weather lenges and require tailored vessel design solutions (Fig. 7).
conditions are required. Rotating equipment such as pumps, power Exploiting NG reserves using LNG technologies in high lati-
generators, and refrigerant gas turbine and compressor units will tudes is commercially viable today at some locations. However, in
require heated and ventilated buildings to house them. Plant layouts more extreme Arctic conditions, new technologies and plant con-
should facilitate easy access to equipment by maintenance staff so figurations must be developed for field development, liquefaction
that both routine maintenance and emergency responses can be and shipping segments of the supply chain. These solutions will
conducted in a safe and timely manner. In fact, plant and equip- be more costly to develop, construct, install and operate than for
ment access under extreme weather conditions need careful con- lower-latitude routes. The LNG industry has the optimism and
sideration. Compressors, pumps, valves, air coolers, wellheads, etc., track record for innovation to justify that acceptable technologi-
require sheltered containment that facilitates easy access and enables cal solutions can be found. Questions, however, remain over the
both staff and equipment to withstand extreme conditions. magnitude of gas reserves yet-to-be discovered and the long-term
sustainability of such high-cost supply chains of natural gas. HP
Arctic LNG shipping. The first ice-class LNG vessels are about LITERATURE CITED
to enter service for the Sakhalin-II project in eastern Russia. Five 1 “United Nations Convention on the Law of the Sea,” Dec. 10, 1982, Annex
new LNG ships will service the liquefaction terminal at Prigorod- 2; Article 4.
2 Latham, A., “Arctic has less oil than earlier estimated,” Oil & Gas Journal,
noye in Aniva Bay. Three were built in Japan with the Moss-type
independent tank and hulls designed to Finnish-Swedish ice-class Nov. 13, 2006.
3 Laherrere, J., “Arctic Oil and Gas Ultimates,” The Oil Drum, March 11,
1B standard; two ships were built in South Korea, each with dif- 2008, http://europe.theoildrum.com/node/3666.
ferent membrane tank designs. All five ships have their propeller 4 ABS, press release: “First Joint Rules for LNG Class Societies ABS and RS
and line shafting built to the Russian Maritime Register of Ship- Jointly Develop Rules for Arctic Gas Carriers,” April 10, 2008
5 Martinez, B., S., Huang, C. McMullen and P. Shah, “Meeting Challenges of
ping ice-class LU2 standard and membrane containment ships
Large LNG Projects in Arctic Regions,” 86th Annual GPA Convention, San
also have their ice-strengthened hulls built to that standard.7 The Antonio, March 11–14, 2007.
performance of these vessels will provide an indication of the 6 Omori, H., H. Konishi, S. A. Ray, F. F. de la Vega and C. A. Durr, “A new
standards required for a more extensive Arctic LNG carrier fleet tool—efficient and accurate for LNG plant design and debottlenecking,”
to withstand sea ice seasons of 100 days and more. LNG, 13, Seoul, 2001.
7 Tustin, R., “From Russia with LNG,” Ice Focus (Lloyd’s Register), April 2006.
As LNG supply chains develop, it is not just at the liquefaction ter- 8 Scherz, D. B., “Arctic LNG: Keys to Development,” 6th Annual LNG
minals where sea ice will be encountered. Plans to build regasification Economics and Technology Conference, Houston, Jan. 30–31, 2006.
9 UNEP/GRID-Arendal Maps and Graphics Library, 2005, http://maps.grida.
terminals along the St. Lawrence River in Canada suggest that the
ships may have to operate in ice at both ends of their routes. The no/go/graphic/major-and-minor-settlements-in-the-circumpolar-arctic.
10 CRUTEM3v dataset. Climate Research Unit, University of East Anglia, June
power installed and the ice class of the vessels apply to the more chal- 2007, http://www.cru.uea.ac.uk/cru/data/temperature, In UNEP/GRID-
lenging Arctic routes, such as to the Western Arctic coastline of Rus- Arendal Maps and Graphics Library, http://maps.grida.no/go/graphic/trends-
sia. They will need to be higher unless dedicated ice-breaker vessels are in-arctic-temperature-1880–2006.
11 Cambridge, UK: Cambridge University Press, “Projected changes in Arctic
commissioned to assist these vessels. With winterization features, such
pack ice (sea ice minimum extent),” In UNEP/GRID-Arendal Maps and
as low-temperature-proof materials to the deck equipment on the Graphics Library, http://maps.grida.no/go/graphic/projected-changes-in-arc-
vessels and on loading and unloading facilities, the ships will have to tic-pack-ice-sea-ice-minimum-extent, 2007.
12 Arctic Climate Impact Assessment (ACIA), 2004, “Shift in climatic zones,
withstand severe wave conditions and persistent cold environments.
Carriers using membrane-containment designs will need reinforced Arctic scenario,” In UNEP/GRID-Arendal Maps and Graphics Library, http://
maps.grida.no/go/graphic/shift-in-climatic-zones-arctic-scenario, 2007.
tank supports to avoid cargo-sloshing damage. Indeed, membrane
designs will need to prove their reliability under such challenging con-
ditions before operators will order them for Arctic service. LNG ships David Wood is an international energy consultant specializing
built for dedicated service to the SnØhvit LNG facility in Northern in the integration of technical, economic, risk and strategic infor-
mation to aid portfolio evaluation and management decisions. He
Norway (ice-free all year) are all of the Moss-type design. holds a PhD from Imperial College, London. Research and training
The challenges associated with first-year ice navigation and concerning a wide range of energy-related topics, including project
those with multi-year ice navigation are very different. Multi-year contracts, economics, gas/LNG/gas-to-liquids, portfolio and risk analysis are key parts
ice is prevalent in the Kara Sea and for year-round navigation with of his work. He is based in Lincoln, UK, and operates worldwide.
icebreaker assistance. Typical hull-structure design values over ice-
sheet thicknesses vary from 120 cm to 170 cm in the summer and Saeid Mokhatab is a consultant for XGAS Ltd, Canada. His
autumn seasons and 170-cm to 320-cm thickness (with hummocks) principal interests include gas engineering, with particular empha-
in the winter and spring seasons.7 Movement in such winter condi- sis on natural gas transportation, LNG, CNG and processing. He
tions requires very powerful engines (85 MW to 120 MW), nar- has participated in several international gas-engineering projects
and published over 180 technical papers and magazine articles
rower beams and strong propulsion equipment to push ice-breaking as well as the Elsevier’ Handbook of Natural Gas Transmission & Processing, which
hulls that are moving slowly (2 nm/hr).8 Although the highest ice- has been well received by the industry and academia. He is the co-editor-in-chief of
classed LNG vessels do need to have ice-breaker assistance at times, the Elsevier’ Journal of Natural Gas Science & Engineering as well as a member of
the vessels and support services will not only be expensive, but the the editorial boards for most of professional oil and gas engineering journals, and
serves on various SPE and ASME technical committees. He served on the Board of SPE
periodic slow speeds along the most challenging parts of their routes London Section during 2003-5, and was a recipient of the 2006 SPE Editorial Review
will require more tankers to transport similar contract quantities Committee’ Technical Editor Awards.
58
In-line laboratory and real-time

quality management
An in-depth look at NIR spectroscopy
M. VALLEUR, Technip, Paris, France
P
rocess plants have traditionally relied on laboratory-quality • Nondestructive methods
determinations and a limited number of in-line measure- • Very fast answers, about 10 to 200 times faster than ASTM
ments to control feed qualities, intermediate streams and methods for some quality determinations, such as octane, cetane,
commercial products. Driven by a very demanding economic detailed hydrocarbon analysis or crude true boiling point (TBP)
environment, this situation has changed dramatically with prog- • Fiber optic use provides a safety advantage in oil refineries and
ress in reliable, accurate and affordable process spectrometers, the possibility for fast multiplexing on several process streams
advances in spectral information processing techniques (chemo- • Easily maintained.
metrics) and availability of fast real-time computers. MIR offers the most sensitive spectra in the 2,500–20,000-nm
Spectroscopic methods have found applications in many sec- domain with a “fingerprint” region between 5,000–15,000-nm
tors, including agricultural and environmental sciences, food where functional absorption bands can be related to organic func-
and beverage, the pharmaceutical industry, electronics, oil and tional groups and be used for quantitative analysis of an individual
gas, petrochemicals, etc. Refer to Workman’s article for a more component. This is the case for cetane booster additives used in
comprehensive review of applied spectroscopy in the infrared gasoil blending. However, the strong absorption requires extremely
domain.1 Applications in the process plants essentially relate costly fiber optics and very short optical paths, making MIR spec-
to oil refining, chemicals and petrochemicals, and impact the troscopy economically difficult to justify for in-line use.
economics and operation organization. NIR has become the favored spectroscopic method in the oil
Since spectroscopy allows for a deep knowledge of chemical industry due to its robustness, high photometric and wavelength
entities, the methods have enabled a number of advanced process accuracy, and short response time compared to the traditional
control (APC) and real-time optimization (RTO) applications ASTM methods.3 Operating at shorter wavelengths, the energy
that could not be achieved with traditional analytical methods level is higher and provides better signal/noise ratio than MIR.
for cost and process dynamic reasons. However, NIR spectra are made of broad absorption bands that
require extensive mathematical processing to extract meaningful
Process plant spectroscopic methods. Most process quality information.
plant laboratories are using several spectroscopic methods, includ-
ing ultraviolet (UV), visible (VIS), near-infrared (NIR), fluo- NIR principles. NIR spectroscopy operates in the 780–2,500-
rescence X, etc. There has been much debate on the compared nm (12,800–4,000 cm–1) electromagnetic spectrum regions,
merits of each method and Chung’s article gives a more detailed consult Workman’s article for a basic introduction to NIR.4 Any
description.2 It appears that nuclear magnetic resonance (NMR) molecule having C-H, C-S, C-N or O-H bonds can be analyzed
and mass spectrometry, although both are powerful and sensitive by NIR. First, second and third overtones are to be found in the
methods, are difficult to implement and maintain online in an 800–2,000-nm domain while combinations give absorption bands
industrial environment due to the high-level skills required. in the 2,000–2,500-nm domain. Low intensity and broad overlaps
require very low signal/noise factors from accurate spectrometers.
Raman spectroscopy has specific merits and has been used
successfully in BTX (benzene, toluene and xylene) plants. Some NIR spectrometer use for industrial applications.
advantages of Raman spectroscopy are: The complex analysis of NIR spectra became feasible when fast
• Fine analysis of chemical mixtures, including isomers computers were made available along with powerful chemometrics
• No requirement to remove water from sample software, efficient detectors and affordable fiber optics. NIR is the
• True simultaneous detectors, no beam splitter required most versatile spectroscopic method with at least 15,000 papers
• Frequency ranges close to visible, allowing the use of inex- published on the technology fundamentals and applications.
pensive long optical fibers (up to 350 m).
With NIR and MIR spectroscopy, experience has shown that Chemometrics. Useful information extracted from NIR spec-
vibrational spectroscopy in the NIR and the mid-infrared (MIR) tra is performed by mathematical processing, generally using
domain was the most appropriate technique for online quality statistical techniques. The most commonly used method is partial
determinations, for the following reasons: least squares (PLS) and its derivatives combined with principal
I 59
components analysis (PCA). Although widely available, it has Refinery process units. NIR applications for quality petro-
severe limitations for complex applications such as blending. leum product determinations were initiated in the US during
Some severe limitations are: World War II. With the contribution of such pioneers as the BP
• Lack of explanation in outlier cases Lavera Research center, these online applications now cover major
• Limited prediction capability for global quality determina- refinery processes such as:
tions, particularly cold properties of gasoil • Atmospheric distillation unit: crude mix true boiling point
• Necessity to calibrate one separate model for each quality (TBP), side stream qualities (naphtha to heavy gasoil)
determination. • Vacuum distillation unit: vacuum gasoil
PLS models may require spectral range optimization to be effec- • Vacuum residue hydrodesulfurization: gasoil, naphtha
tive5 and avoid artifacts from over fitting. Furthermore, they are dif- • Naphtha hydrotreater
ficult to transfer from one spectrometer to another. They are widely • Hydrodesulfurization gasoil, wild naphtha
supported by several software technologies and affordable. Also, they • Reformer: feed and reformate
can be efficient on simple applications such as octane on a reformate • Gasoline hydrogenation: gasoline
or alkylate stream and used for fast product identification.6 • Isomerization: isomerate
A more advanced method makes use of topology-based data • Alkylation: alkylate
mining from a spectra reference library. It is proven highly effec- • Aromatics units: feed and BTX extract
tive on very complex NIR applications. The specific advantages • FCC unit: feed, light gasoline, heavy gasoline, light cycle
of this method are: oil, heavy cycle oil
• Uses the whole spectrum of information, including the com- • Hydrocracker unit: gasoline, jet fuel and middle distillates
binations domain (this depends on the optical fiber type used) • Lube oil units: intermediate streams.
• Provides a sample classification by chemical species, a use- More recently, NIR has been used on crude distillation units
ful feature with outliers (unrecognized spectra), that gives a to predict the crude mix TBP (12 distillation points ASTM
physical explanation D2892) in real-time to minimize transient operations dur-
• Allows computation of blending indices for non-linear ing crude swings.8,9 This application is most useful to increase
properties, used in linear programming (LP) models and creates throughput in European refineries processing a large crude slate
virtual blends for the spectral database densification, as shown with frequent swings, sometimes once a day.
in Figs. 1 and 2.
• Predicts responses to some additives Blending. Early NIR applications were quite simple, measur-
• Cumulates spectral information over time, improving pre- ing the reformate octane number, but were quickly extended to
dictions and only requires a single model for all properties of a include very complex gasoline and middle distillates blending.
given process stream. This blending operation is critical as it is the last processing step
Besides the ability to provide the required precision and before selling the commercial product. It also requires accurate
accuracy for quality determinations, the main criterion for the quality determinations for specifications that include the quality
chemometrics selection method allows refinery laboratory staff to certificate for commercial transactions. Tables 1 and 2 provide
maintain NIR models independently on the long-term.7 a quality specifications list that is routinely predicted by NIR
for gasoline and gasoil optimal blending with repeatability and
Oil and gas production. NIR has only recently been used reproducibility equal to or better than ASTM.
to monitor crude production from various gathering centers An NIR-based blending application is performed with
to predict composition at receiving terminals. Given untreated increased efficiency compared to traditional methods.10,11 How-
crude conditions, i.e., sand, sediments and water, the sampling ever, a number of quality determinations illustrated in Tables 3
system is the most critical application. There are on-going proj- and 4 may be required on commercial quality certificates but are
ects to use NIR to determine condensate qualities on gas fields not achievable by NIR or not yet proven.
with an objective to deliver a constant commercial product at It should be noted that:
the loading facilities. • Water in samples can be noticed by NIR but is a nuisance
FIG. 1 Spectral database before primary densification. FIG. 2 Spectral database after MC primary densification.
60
TABLE 1. Gasoline quality determinations by NIR TABLE 3. Some required gasoline quality
determinations
ASTM
Quality determination Unit methods Specification Note Quality determination Unit ASTM methods Specification
Research octane number D2699 Min Water content mg/kg D1744, D1364 Max
Motor octane number D2700 Min Washed gums content mg/100 ml D381 Max
Density Kg/liter D1298 Range 1 Potential gums mg/100 ml D873 Max
Temperature 10% distilled °C D86 Max Oxidation stability minutes D525 Min
Temperature 50% distilled °C D86 Range Copper corrosion D130
Temperature 90% distilled °C D86 Range Doctor test D4952
Temperature FBP °C D86 Max Mercaptan sulfur unit mass % D3227
Reid vapor pressure @ 100°F Psi D323 B, Max 2 Color D1500
D5482
Benzene content % Vol. D6293, Max 3
D5134 TABLE 4. Some required gasoil quality determinations
Total aromatics content % Vol. D4420, Max Quality determination Unit ASTM methods Specification
D1319,
Water and sediments content % Vol D1796 Max
D6293
Water content % Vol D2709 Max
Olefins contents % Vol. D1319, Max 3
D6293 Ashes % Weight D482 Range
Note 1: ASTM D4052 repeatability cannot be achieved by NIR. Lubricity at 60°C Micron ISO 12156-1 Max
Note 2: If no C3 variations.
Note 3: If C > 0.5 % mol. Total acidity mg KOH/g D974 Max
Conductivity pS/m D2624 Min
TABLE 2. Gasoil quality determinations by NIR Copper strip D130
ASTM Total contamination mg/kg D2276 Max
Quality determination Unit methods Specification Note * Total acidity and lubricity are likely to be predicted by NIR.
Cetane number D613 Min
detailed hydrocarbon analysis is performed at NIR spectra acqui-
Cetane index D4737 Min
sition speed and processing, i.e. about once a minute, 200 times
Flash point (PMCC) °C D93 Min
faster than gas chromotography-based methods. Pyrolysis gasoline
CFPP °C D6371 Max partial hydrogenation is optimized using real-time dienes measure-
Pour point °C D6749, D2500 Max ment content. NIR has also been used to determine the ethylene
Cloud point °C D5773, D2500 Max content in flakes or propylene/ethylene copolymer pellets.14
Density @ 15°C Kg/ liter D1298 Range 1
Temperature 90% distilled °C D86 Report
Laboratory methods. Because NIR is a secondary method,
it relies on proper quality determinations on the laboratory spec-
Temperature 95% distilled °C D86 Max
trometer with traditional instruments. Prior to any NIR project, it
FBP °C D86 Report is recommended to certify the laboratory to ensure that best prac-
Kinematic viscosity @ 100°F cSt D445 Range 4 tices are used. Particular care must be given to regular instrument
Conradson Carbon Residue % Weight D4530, Max calibration, sampling procedures and sample conditioning (water
D189 content, for instance), and spectrometer cell temperature control.
Aromatics content % mass D5186, Max
D2429, Spectrometers. The advantages of Fourier transform infrared
D5292 spectrometers (FTIR) have been recognized by process plants, in
Polycyclic aromatics (PAH) % Weight D5186, Max particular repeatability, robustness (no moving parts) and stabil-
D2429, ity. They offer a very high signal/noise ratio.
D5292 FTIR spectrometers performances are brilliant, typically:
Note 1: ASTM D4052 repeatability cannot be achieved by NIR.
Note 4: Without ASTM repeatability.
• Maximum spectral resolution better than 2 nm
• Wavelength accuracy: better than 0.3 nm
for spectra quality • Wavelength repeatability: 0.01 nm
• Gums and oxidation stability are presently indicated by NIR • Cell path length: 500 ± 15 μm
• The traditional copper corrosion and doctor test are not • Absorbance repeatability: 5.10–4
critical with low sulfur gasoline. • Baseline stability better than 1.10–3.
Calibration transfers between laboratory and process spec-
Petrochemical plants. Spectroscopic methods have been used trometers are easily achieved, provided precautions have been
on BTX units and ethylene plants.12,13 Liquid feeds to steam crack- taken on identical cell reference temperature and optical path.
ers are excellent candidates for NIR-based high frequency analysis
to predict PINA by carbon atom and cracking yields to manipulate Sampling systems. Extractive sampling systems are generally
in real-time the cracking furnace severity and adapt to the cold sec- preferred to in-situ probes for complex applications as they allow
tion operating conditions. As for crude TBP determination, this a strict temperature cell control. In-situ probes are essentially used
I 61
in the chemical industry on simple streams that are not subject

to temperature variations and are free of water and solids. Sam-
ple conditioning, such as filtering or water removal is generally
required on oil refinery process streams. Sampling systems can
become quite complex, as shown in Fig. 3, and be a weak NIR
system component from a reliability view point. Together with
the shelters, they are a major CAPEX item, considering sample
extraction, fast loops and sample recovery system. Sampling sys-
tems must also include the reference control and wash chemicals,
generally high-purity toluene and n-Hexane.
Fiber optics. Process spectrometers are frequently multiplexed

on several detectors using fiber optics. Silica-grade fibers used for
telecommunications cannot be used in the combinations domain
because of their high absorption and must be replaced by more
expensive zirconium fluoride grades.
Limitations on sensitivity. Since NIR is not a sensitive

method, it is necessary to use standard ASTM analyzers for the FIG. 3 Sampling system.
following quality determinations:
• Densimeter to obtain ASTM 4052 repeatability
• Gas chromtography or other methods for low concentra- As a consequence, the spectral database population and den-
tions (less than 0.5%), e. g., very low benzene or olefins content sification is the most critical NIR project step, as it must cover
• Sulfurimeter for very low sulfur content such events as:
• Reid vapor pressure (RVP) analyzer if C3 concentration • Crude swings
in the C4 gasoline blending component is subject to significant • New crude imports
variations. • Process unit operating modes
• New intermediate stream imports
Repeatability and reproducibility. Repeatability is • Blend recipe variations
important for advanced process control strategies, as when satu- • Additive changes
rating constraints. Reproducibility is the main performance • Partial process unit shutdowns
indicator when measuring commercial product quality that • Catalyst activity changes
might be re-tested by a third party. In both cases, performance • Seasonal product specifications.
guarantees must not only be agreed upon prior to NIR project There is a significant initial workload for the refinery labora-
signatures on both repeatability and reproducibility but also on tory to achieve the required database density, but when the models
the acceptable outlier ratio, measuring the NIR model robust- are properly calibrated and maintained, NIR can provide superior
ness. The NIR model robustness is the most difficult issue—any results, for example on gasoil blending as illustrated in Table 5.
condition that impacts the chemical species must be taken into
account to avoid outliers. System integration. To capture all its benefits, NIR applica-
tions require a strong integration with many other sub-systems
and they are:
TABLE 5. NIR vs ASTM reproducibility results
• Distributed control system
ASTM NIR ASTM • Laboratory information management system
Quality determination method reproducibility reproducibility • Advanced process control
Cetane number D613 1.9 4.0 • Real-time optimization
Cloud point D2500 2.7 4.0 • Instrumentation maintenance
CFPP D6371 2.5 3.5 • Analyzer data validation system.
IBP D86 7.7 8.5
Plant acceptance. Since it impacts the responsibility matrix
E 95 D86 5.5 8.5
between laboratory and maintenance, implementing NIR in a
E 250 D86 2.3 6.2 plant is not straightforward.
E 350 D86 1.5 3.2 The main acceptance criterion is conformance with primary
E 360 D86 1.4 1.5 standards, essentially ASTM and ISO. This must be observed
FBP D86 4.2 10.5 over a time period, typically six months, to make sure the repro-
Flash point D93 3.6 5.0
ducibility is not affected by operating conditions and seasonal
change of transportation fuel specifications. NIR models should
Viscosity ISO 3104 0.06 0.05
never be accepted on the basis of calibration statistics that ignore
Poly aromatics IP391 0.2 1.8 the practical operation range.15 Another fundamental prerequi-
Aromatics IP391 0.3 4.4 site to success is to find an NIR champion within the laboratory
Specific gravity D4052 1 0.5 staff to not only be the focal point but also to implement the
necessary changes to the work processes.
62
Maintenance burden. Maintaining FTIR spectrometers heated cell. Refineries have also tested NIR use to predict the
is very easy compared to traditional ASTM analyzers. Designed bitumen penetration quality.17 More recently, new techniques
originally for space missions, the hardware is extremely robust. based on automatic solvent dilution have been implemented on
Unfortunately, sampling systems still require attention as they a laboratory spectrometer at line to provide quality heavy feed
are likely to plug and/or leak. The most critical task is the NIR determinations, such as vacuum residues.18 Quality determina-
models maintenance burden. The plant laboratory must be tions for FCC feeds typically include: density, Conradson carbon
able to absorb the workload of expanding the spectral database residue, sulfur, total acid number, basic nitrogen, distillation
and taking care of outliers. Lacking model support is the first curve, detailed aromatics analysis and viscosity. Compared to
NIR project failure cause, followed by indefinite re-modeling traditional laboratory analysis, NIR has a significant advantage
(generally due to inadequate chemometrics) and poor reliability by updating at high frequency the quality determinations that
of sampling systems.16 OPEX under-estimation related to NIR are required by APC and RTO. There is ongoing developmental
models maintenance is the shortest route to project failure. work to predict bitumen quality determinations.
NIR advanced applications. The fol-

lowing are some applications that can bring
additional benefits.
Blend indices. NIR spectra contain the
non-linearity information for such proper-
ties as RVP, flash point, distillation points,
octane, cetane, cold properties and viscos-
OUR MISSION
ity. Therefore, they are used to predict the
blend indices to be used in LP models,
to correct blending recipes, taking into
YOUR SAFETY
account heels and to feed-forward real-
time optimal control—all very useful for
in-line certification.
In-line certification. When logistics are GET MORE
tight, there is a strong interest for loading
products directly from the blender header to
a sea tanker without the need to fill a refin-
WITH HART
ery tank, isolate, sample, analyze and then MSA Ultima® X Series
release. This in-line certification process Gas Monitors
now available with HART Protocol.
requires accurate, fast and reliable online
quality determinations, exactly what NIR • More efficient
is providing. The majority of new grassroot asset management
refineries being built in the Middle East • More flexibility with digital
and Asia are planning to use this efficient or analog capability
procedure. • More compatibility with
Additives management. Many additives existing installed operations
are used in the oil refining industry, in gasoil
Ask about our new 10-year warranty
blending, and may include:
on DuraSource™ Technology for
• Cetane booster Ultima XIR and XI Gas Monitors.
• Cloud-point depressants For your gas detection solutions,
• Flow improver (MDFI) contact MSA at 1.800.MSA.INST.
a
• Drag reducing agent
• Lubricity improver VISIT US ONLINE
• Anti-static
• Oxydation stability
MSANET.COM
• Wax anti settling Visit us at ISA
• Corrosion inhibitor Booth #2350
• Bactericide
• Anti-foam.
Presently, NIR provides cetane-booster
and cold-property additive responses.
Using combined NIR and MIR offers a
large potential for optimized additive dos-
| G A S M O N I T O R S | S C B A | M U LT I G A S D E T E C T O R S |
age, a significant operating cost savings.
| HEAD/EYE/FACE PROTECTION |
Heavy process streams. Early work on
quality determinations of heavy streams by 1.800.MSA.INST | www.MSANET.com/hydrocarbon.html
NIR started with FCC feeds on a labora-
tory FTIR spectrometer equipped with a
I 63
Additional quality determinations. Research is underway Point, viscosity index (VI), viscosity
to extend the range and quality determination accuracy, such as • Dewaxing unit: oil content and slack wax viscosity
gasoil viscosity, presently on the borderline of ASTM reproduc- • Furfural unit: % PCA, viscosity and % S extract, % PCA,
ibility. Potential gums and oxidation stability should be acces- VI of raffinate
sible by NIR, at least for indication. Gasoil lubricity is becoming • Deoiling unit: wax oil content.
a constraint with very low sulfur gasoil and could benefit from
NIR in-line determination with additives. Fuel oil and bitumen In-line laboratory. New refineries are becoming very com-
blending could be optimized using NIR. plex in terms of process unit numbers, sometimes over 50. In
Lube-oil characterization on laboratory spectrometers has addition, the crude slate can be extremely wide in European
proven feasible and could be extended for on-line use in APC refineries. The new export refineries in the Middle East and India
strategies, in particular, the following units: will produce a very wide range of commercial grades, including
• Hydrofinishing unit: % PCA, Conradson carbon, Pour up to 15 different grades of gasoil. Quality determination num-
bers requested by process unit and blending
operations are growing significantly. Table
6 illustrates quality determinations on a
laboratory FTIR spectrometer for commer-
Mission: Immersion. cial products of an export refinery in the
Middle East.
Immersion Engineering ™
™
This is an incentive to systematically
goes deep to solve your use in-line NIR spectrometers to obtain
high-frequency quality determinations at
heat transfer problems. acceptable CAPEX and OPEX.
One spectrometer can analyze several
streams:
• Streams can be multiplexed optically
on multi-channel FTIR spectrometers
whenever a high frequency of data acquisi-
tion is required, e.g., an APC application
with high dynamics.
• Liquid multiplexing by the sampling
system can be used when the stream quali-
ties are not critical.
• In practice, a mix of two types of multi-
plexing is implemented on one spectrometer,
providing quality determinations on as many
as 16 streams with frequencies between less
than 1 minute and 15 minutes.
Each stream has between 5 and 10 qual-
ity determinations, so one spectrometer can
deliver between 80 and 160 quality deter-
minations. If four or five FTIR spectrom-
eters (depending on plant topology) are
strategically placed in a refinery, between
Even though you may call us on the performance and applications of heat 300 and 600 quality determinations are
phone miles away, we're so deep into transfer fluids than we do. available online, justifying the label “online
your stuff--your fluid, your equipment, So pick a service and call one of our laboratory.”
your system--we can virtually touch it, technical specialists. Or, check out our
see it.
In a recent front end engineering design
web site for case histories, data sheets,
Immersion Engineering is a bundle of comparisons, user’s guide, tip sheets (FEED) for a grassroot refinery in the Middle
very specialized services that you can and technical reports. It’s all there, it’s East, NIR systems were designed to be used
cherry pick. Some are free, some you deep, it’s Immersion Engineering. on the 30 streams, as shown in Table 7.
pay for. We’re the only company offering NIR spectrometers have been used so
them all. far on liquid streams. More recently, NIR
One thing is for sure; when you need tunable diode laser analyzers are being
HTF help you need it now. Nobody applied for quality determinations, includ-
knows more about the chemistry, HEAT TRANSFER FLUIDS
ing traces on gas streams.19 The primary
4 Portland Road ®
applications in gas processing or liquified

Eyeball this selection of services. West Conshohocken PA 19428 USA
natural gas plants are for moisture analysis,
800-222-3611
®
■ Fluid Analysis ■ Troubleshooting H2S, CO2, NH3 or HCl but TDL could
610-941-4900 • Fax: 610-941-9191
■ Fluid Maintenance ■ System Layout be used in refinery gas plants, extending
[email protected]
■ Training the online laboratory range with very fast
www.paratherm.com
and sensitive quality determinations.
64
GAS PROCESSING DEVELOPMENTS
TABLE 6. FTIR spectrometer quality Spectroscopy Europe 13/2, pp. 10–14,

determinations for commercial products 2001.
8 Park, J., K. E. Kim, I. Cho, Use
Product Grades Determinations Specifications of real-time NIR spectroscopy for

the on-line optimization of a crude
Mogas 4 28 7 distillation unit, NPRA 2000 com-
Jet kero 8 88 11 puter conference, CC-00-159.
9 Sela, I., N. Fontjin and I. Zilberman,
Condensate 1 6 6
Gasoil 16 116 11
“Software speeds implementation of
analyzer in crude unit,” Oil and Gas
Journal, April 10, 2000.
Gas Processing
Naphtha 6 54 9
Total 35 292
10 Vötsch, R., and M. Valleur, “Einsatz
der NIR technologie beim in-line

Engineers and
TABLE 7. NIR for a grassroot refinery
blending von Ottokraftstoff,“ Erdöl
Erdgas Kohle, 114 Jahrgang, Heft 6, Other Industry
Category Streams Determinations Comment
June 1998.
11 Barsamian, A., “Get the most out of
your NIR analyzers,” Hydrocarbon

Professionals
Mogas pool 7 78 Processing, January 2001.
12 Ku, M., H. Chung and J. Lee,
You Know
Middle distillates pool 6 84
Process units 13 127 “Rapid compositional analysis of ■ Heavy hydrocarbon and water
Heavy streams 4 28 At line
naphtha by Near-Infrared spectros- in natural gas may form
copy,” Bull. Korean Chem Soc., Vol.
19, No. 11, 1998.
condensate in export lines.
Total 30 317
13 Lambert, D., B. Descales, S. Bages, ■ Wet (water saturated) natural
S. Bellet, J. R. Llinas, M. Loublier, gas may form hydrates

Extended use of spectroscopic methods J. P. Maury and A. Martens, “Optimize steam and plug equipment and
in process plants has been made feasible by cracking with online NIR analysis,” Hydrocarbon
transportation lines.
the availability of robust and affordable 14 Processing, December 1995. ■ Maul operation of a compressor
Barnes, S. E., M. G. Sibley, H. G. M. Edwards
hardware and powerful mathematical pro-
cessing of spectral information. New appli-
and P. D. Coates, “Applications of process spec- may result in “surge” and
trometry to polymer melt processing,” Spectroscopy
cations are being developed in all sectors Europe, 15/5, 2003. “stone wall.” Surge may
of the oil and gas industry, allowing real- 15 Davies, A. M. C. and T. Fearn, “Back to basics: destroy a compressor.
time quality control from feeds receipts to calibration statistics,” Spectroscopy Europe, Vol. 18,
But Do You Know
products liftings. The concept of “in-line 16 No. 2, pp. 31–32, 2006.
Barsamian, A., “Optimize fuels blending ■ How to avoid condensation
laboratory” is becoming a reality. HP with advanced online analyzers,” Hydrocarbon
Processing, September 2008. from forming?
ACKNOWLEDGMENT 17 Blanco, M., S. Maspoch, I. Villarroya, X. Peralta, ■ How to prevent hydrate
J.M. Gonzalez and J. Torres, Analyst, Vol. 125, pp.
This article was revised and updated from an
earlier presentation at the NPRA 2008 Plant Auto- 1823–1828, 2000.
formation and plugging of
mation Q&A and technology meeting in Orlando, 18 Lambert, D., C. St. Martin, M. Sanchez, B. equipment and pipelines?
Florida. Ribero and S. Beauchamp, “FCC Heavy feed ■ How to safely operate and
characterization for process control through
TOPNIR analysis,” ARTC Conference, March
protect compressors from
LITERATURE CITED
1 Workman, J., “Review of process and non-
2008. surge and stone wall?
19 Miller, S., “TDL technology promises improved
invasive Near-Infrared and Infrared spectroscopy:
process control in gas plants,” Gases & Take the Campbell Gas CourseTM
1993-1999,” Applied Spectroscopy Reviews, Vol. 34
(1&2), pp. 1–89, 1999.
Instrumentation, March/April 2008. (G-4 Gas Conditioning and
2 Chung, H. and M. Ku, “Comparison of Near- Processing) to learn these
Infrared, Infrared and Raman Spectroscopy for answers and more.
the Analysis of Heavy Petroleum Products,” Marc Valleur is the manager of
Applied Spectroscopy, Vol. 54, No. 2, 2000.
3 Davies, T., “The history of near infrared spectro-
the Advanced Systems Engineering For a list of G-4 course
(ASE) business line of Technip France.
scopic analysis: past, present and future,” Analusis He has over 30 years of experience dates and locations go to
Magazine, Vol. 26, No. 4, M17-M19, 1986. with large, multinational control and www.jmcampbell.com/HCP
4 Workman, J., “An introduction to Near-Infrared information systems for the oil, gas and petrochemicals
Spectroscopy,” Spectroscopynow.com, March industries in managerial and senior consultant posi- For a FREE subscription to the
2004. tions. His technical fields of expertise include database Campbell Tip of the Month go
5 Lee, Y., H. Chung and N. Kim, “Spectral range management systems, unattended operations, oilfield
to www.jmcampbell.com/TIP2
optimization for the near-infrared quantitative and process plant integrated decision support systems,
analysis of petrochemical and petroleum prod- advanced process control and near-infrared technol-
ucts: naphta and gasoline,” Applied Spectroscopy, ogy, blending reengineering and offsite operations.
Mr. Valleur is an expert for the EEC on computerized
Vol. 60, No. 8, pp. 892–897, 2006.
6 Chung, H., Hyuk-Jin and M. Ku, “Rapid identi- energy management systems and the Technical Assis-
tance to the Commonwealth of Independent States
fication of Petroleum Products by Near-Infrared (TACIS) program. He is also an associate professor at
Spectroscopy,” Bull. Korean Chem. Soc., Vol. 20, the French Petroleum Institute (ENSPM-FI). He holds
No. 9, 1999. an MSc degree in chemistry from the Paris University
7 Fearn, T., “Chemometrics for near-infrared
(ENSCP) and specialized in chemical engineering at
spectroscopy: past, present and future,” Institut Français du Petrolea.

MAINTENANCE/ROTATING EQUIPMENT
Auxiliary pumps and support

systems for process machinery
Proper system design and operation are critical
to plant uptime and reliability
J. R. BRENNAN, Colfax Corp., Monroe, North Carolina
M
uch has been written on the subjects of process pumps, with American Petroleum Institute (API) Standard 614: Lubrica-
pipeline pumps and similar mainstream hydrocarbon tion, Shaft-Sealing and Oil-Control Systems and Auxiliaries.
processing machinery. Such equipment is obviously Probably the simplest systems are found on gear-speed reducers
critical to ongoing operations, but little has been produced cover- (or increasers) and large centrifugal pumps. The main oil pump
ing auxiliary support pumps. is frequently machine-driven, whereas the standby pump is most
Most rotating machinery within refineries, petrochemical com- commonly electric-motor-driven.
plexes and chemical processing plants requires forced lubrication, The standby pump is started before the machine. Once the
and many process-gas centrifugal compressors and turboexpand- machine is up to normal running speed range, the standby pump
ers also require forced-oil systems to seal process gas within the is shut down and remains in standby mode. Should the lube-oil
machine. Control-oil systems are also common, supplying an oil header pressure fall below some setpoint, a pressure switch will
flow proportional to machine speed. Proper system design and cause the standby pump to start. Fig. 1 shows a simplified sche-
operation of these auxiliary pumps is critical to plant uptime matic of this type of lube system.
and reliability, considering 24 months between turnarounds and The standby pump may be of the external, horizontal type or
24-hr-per-day operation are normal. of the vertical, in-tank arrangement. Normally, all external pumps
Rotating machinery large enough to require forced lubrication are steel cased to minimize risk of a lube pump case fracture dur-
will normally have both main and standby lube-oil pumps. These ing a fire, which might allow lube oil to escape in large volumes
can be used for prelubrication before starting the machine, con- that could fuel an otherwise small fire.
tinuous lubrication while the machine is running (even if a lube Fig. 2 shows 1,100-hp twin-screw pipeline pumps. Their tim-
pump, driver or power supply goes down) and lubrication during ing gears and antifriction bearing system are force-cooled and
coast-down, which can take several minutes or more for very large lubricated using a small-flow gear pump driven from the outboard
machine sets. These oil systems are frequently designed by the end of one of the pipeline pump’s rotors. An oil reservoir, filter
machinery manufacturer, and many are constructed in accordance and air-to-oil heat exchanger complete the system.
In this case there is no standby pump
since the pumping station has full standby
twin-screw pump capacity. Note that driv-
Three-way PCV Oil supply ing the auxiliary pump from the machine it
PI Cooler bypass PI TI PSL PSL ¾ in.–150#
serves is the most reliable method to ensure
valve R.F. flange
that power is available to the lube pump.
CWS Filter As long as the machine rotates, the pump
supplies cooling flow. The site location for
Motor Motor Filter these machines, Venezuela, could not readily
pump PSV pump provide cooling water for the lube system;
thus the radiator/fan arrangement of the
Fill/vent heat exchanger.
60-gallon Oil return UNLESS OTHERWISE SPECIFIED
DIMENSIONS ARE IN INCHES ENGINEERING DATE Large rotating equipment trains may
TI reservoir 2 in.–150# need cooling oil flowrates in excess of 1,000
LG TOLERANCES DRAWN BY: JGF 3-13-99
Suction R.F. flange DECIMALS ± 0.0625´ CHECKED BY: WT 4-13-99
strainer
LS FRACTIONS ± 1/16´ APPROVED BY: JGF 4-28-99 gpm, usually at pressures of about 75 to
Heater ANGULAR ± 1*
MATERIAL:
150 psi. Such systems invariably have main
Drain and standby auxiliary pumps. Frequently
1 in. NPT
the main pump is steam-turbine-driven, if
steam is available on site, and the standby
FIG. 1 Simplified lube-oil system schematic. pump would normally be driven by a con-
ventional AC electric motor.
66
FIG. 2 1,100 hp-twin-screw pipeline pumps with integral lube-oil

pump. FIG. 4 Small three-screw lube-oil pump for smaller flow
requirements.
Inlet
Discharge
FIG. 5 Section view of a large-flow double-suction lube-oil pump.
machine train. The installation is at an oil refinery in Jurong,

Singapore.
For cost and efficiency, auxiliary pumps should be sized to
operate at two- or four-pole motor speeds (1,500 to 3,600 rpm),
if possible. This results in less costly pumps and drivers as well as
better pump and driver operating efficiencies. Positive-displace-
ment rotary pumps are usually preferred over centrifugal pumps
for these auxiliary services since they are self priming, do not
FIG. 3 Turbine/compressor train on test with lube-oil console become air bound, have very predictable performance and are
(right).
simple to control.
Fig. 4 is a small three-screw pump typical for lubricating
If two AC motors are used for main and standby service, they smaller rotating machinery. Each wrap of the screw set forms a
should be wired to different power sources so the failure of one chamber, relatively independent of adjacent chambers. Pressure
source does not compromise the machinery train. In some cases, rise across the pump is effectively staged and causes very low
a third “coast-down” pump may be desired, frequently driven internal unit loading.
by a DC motor supplied with power from a trickle-charged Fig. 5 is a section view of a pump similar to those shown in
battery bank. Fig. 3. Flow is split equally at the pump inlet and delivered to the
Fig. 3 is an overhead photograph of a steam-turbine/hydrogen- pump outlet in a smooth, continuous manner.
booster compressor train on test. The lube-oil console is to the Because of the opposed flow pattern in these pumps, internal
right. A pair of double-suction three-screw pumps is at the lower axial hydraulic forces due to differential pressure are canceled.
right, one a motor drive, the other a steam-turbine drive. Each Radial forces are reacted in the hydrodynamic oil films surround-
pump provides about 530 gpm to all the bearings in this four- ing the pumping screws. These features, together with the pres-
I 67
sure-staging effects of the pumping ■ Auxiliary pumps are small but particle sizes around 150 microns
chambers, result in very long oper- or even larger. It is, therefore,
ating life for this kind of pump. important parts of process industry important that new installations be
Most centrifugal process com- thoroughly flushed using separate
pressors and turboexpanders use reliability. Provide them the environ- flushing pumps. Flushing pumps
labyrinth or mechanical shaft seals should be large-clearance centrifu-
to contain the process gas, and these ment that they need and they will gal pumps that can supply system
seals frequently are supplied with provide many years of trouble-free flows higher than the lube pumps
cooled lube oil at a pressure just so the high velocities produced
slightly higher than the gas pressure service. encourage debris to be moved to
at the seal. Depending on the service the filters. Once the system is veri-
involved, the seal-oil pressure can range to 4,000 psi or higher. fied clean, the rotary pumps are ready for use.
Again, three-screw pumps are typically used for this demand- Another system problem more common than it should be is
ing 24x7 service. They are frequently boosted from the lube pump excessive lube-oil aeration. Almost all lube-oil systems gravity
system, sharing the same oil system. Since the pressure demand on drain the lube oil returning to the oil reservoir, which aerates the
many of these pumps is much higher than lube pumps, they will oil during its passage through the machinery served. If the reser-
normally have many wraps or stages (up to 12) to effectively resist voir is not properly baffled, this aerated flow will travel directly to
internal slip through running clearances and maintain internal the auxiliary pump inlet where pressure will be lowered more and
loading at low levels for prolonged operating life. the air content by volume expanded. When that occurs, pump
By far, the most vulnerable time for auxiliary pumps is their ini- operation can become noisy, erratic and result in system shut-
tial startup. The culprit is almost always hard, solid contaminant downs. Severe cases can cause pump damage or destruction.
in the oil system. Because rotary, positive-displacement pumps are To reduce aeration, all oil return lines need to terminate below
close-clearance devices, they do not generally survive well in the the minimum oil level in the reservoir. Baffles need to be arranged
presence of pipe scale, weld bead, metal filings, machining chips within the reservoir to maximize the time that the oil is allowed to
and other debris typical of a new installation. release entrained air before entering the pump again. A 10-min-
While the served machinery oil flow is usually filtered to the ute retention time is a fairly standard reservoir sizing criteria
10-micron range before it reaches critical bearing clearances, (minimum reservoir volume equals 10 times the pump flowrate).
the flow to the pump inlet may go through a strainer that stops Improper or no baffling will defeat the retention time by allowing
return oil to “short circuit” directly back to the pump.
Some rotating machinery will drain lube oil to the machine
sump or, in some cases, the crankcase of a large reciprocating
Improve plant reliability machine. This oil needs to be pumped, rather than gravity-
drained, to the main oil reservoir. Positive-displacement scavenge
with these must-have books pumps (machine- or motor-driven) are used for this service and
are sized to displace about twice the main oil pump’s rated flow.
These pumps deliver about 50% air and 50% oil at very low
pressure (usually 10 to 15 psig), ensuring that lube oil does not
Machinery Failure Analysis accumulate in the machine sump.
Handbook Rotary positive-displacement pumps can also be used as hydrau-
lic power recovery motors (HPRMs). Processes that reduce liquid
Helps anyone involved with machinery pressure by throttling are prime candidates for dropping pressure
reliability to understand why process
equipment fails.
across an HPRM which, in turn, can power a partial-capacity feed
pump or a plant air compressor. Otherwise wasted energy (throt-
www.GulfPub.com/MachFailureAnalysis
www G tling) is recovered at efficiencies up to 75%. Given today’s energy
costs, HPRMs are well worth their expense.
Auxiliary pumps are small but important parts of process
industry reliability. Provide them the environment that they need
and they will provide many years of trouble-free service. HP
Improving Machinery
Reliability
Heinz Bloch provides proven techniques
and procedures that extend machinery
life, reduce maintenance costs and achieve
optimum machinery reliability.
James R. Brennan is a consultant for Colfax Corp. (NYSE:
www.GulfPub.com/ImprovMachReliability CFX), a global leader in critical fluid-handling solutions, including
the manufacture of positive-displacement pumps and valves for
oil & gas, power generation, commercial marine, naval and other
Gulf Publishing Company industrial applications. Located in Monroe, North Carolina, USA, his
+1-713-520-4428
1 713 520 4428 l +1-800-231-6275
1 800 231 6275 responsibilities encompass worldwide technical support and service for Colfax’s Hout-
Email: [email protected] tuin, Imo and Warren brand pumping applications. Mr. Brennan is a 1973 graduate
of Drexel University in Philadelphia, Pennsylvania, USA, a member of the Society of
Petroleum Engineers (SPE) and has 39 years of service with Colfax.

68
PROCESS DEVELOPMENTS
Consider practical conditions

for vacuum unit modeling
A good simulation model is a tool that reveals critical operating
conditions and can be applied to daily operations
R. YAHYAABADI, Esfahan Oil Refining Co., Esfahan, Iran
S
imulation tools are frequently applied to identify critical Preventing coke formation requires sufficient wash-oil flow to
operating conditions. Modeling operating parameters will keep the middle of the packed bed wet; otherwise, high-residence-
help ensure better unit reliability. Some operating parameters time stagnation zones are created.4 Coke forms in the middle
cannot be measured directly. In such cases, the parameters are cal- because it is the only part of the bed that is not wetted.4 Coking in
culated via a model. In a revamp case, simulation models are tools the middle of the wash zone has been discussed in the literature.7–9
used to determine project goals. Too often, revamp projects failed Wash-zone efficiency has a large effect on the HVGO quality. Small
due to incorrect simulations. The author discusses tips to improve changes in the 95 vol% EP distillation tail have a large impact on
simulation methods when revamping crude vacuum units. GO product metals.2 Increasing wash-section efficiency can reduce
the GO product 95 vol% EP distillation tail and metals.2
Vacuum units. Many different types of vacuum towers are used Coking in the heater outlet is a common problem.5 Coke forms
in refineries.1 The typical and most common refinery vacuum inside the radiant section tubes of the vacuum heater, because the
unit is shown in Fig. 1. In this vacuum unit, the feed (atmo- oil film flowing along the inside of the tube exceeds the tempera-
spheric residue—long residue) is separated into two vacuum gasoil ture and residence time needed to initiate thermal cracking.5 So,
products—light vacuum gasoil (LVGO) and heavy vacuum gasoil controlling the oil-film temperature and residence time is essential
(HVGO). Typically, VGOs are sent to catalytic units for further to minimizing coke formation.5
processing (conversion).
The refinery’s main objective is to increase VGOs yield to Vacuum unit design. Vacuum unit design can influence
improve plant profitability. Higher yields mean higher true boiling VGO yield, product quality and run length. 2 When designing
point (TBP) cutpoints. At the same pressure, increasing the TBP
cutpoint allows higher heater outlet and flash-zone temperatures. To vacuum system
For catalytic processes using VGOs, there are some limitations
regarding metal content, microcarbon residue (MCR) and/or
asphaltenes of the feed. In this processing operation, increasing
the TBP cutpoint can be done while minimizing the metal con-
Vacuum LVGO
tent of the LVGO and HVGO. Process and equipment designs column
that minimize the distillation tail will reduce metals.2 Minimiz-
ing HVGO metals will dramatically increase catalyst life.3 This
problem could become critical, especially for HVGO.
HVGO
Vacuum unit critical operating conditions. The most
common important problem of vacuum units is coke formation in Feed
Wash oil
fired heater and wash sections. This is a matter that has been dis- Wash zone
cussed in many articles. Wash-bed coking continues to be a common Collector tray
problem affecting vacuum unit run length.4 In several cases, vacuum Vapor horn
Fired Flash
heater and column wash sections coked in less than one year.5 heater
Transfer
zone Slop wax
line
Wash zones continue to coke causing poor HVGO product
quality, low HVGO yield and unscheduled outages to replace pack-
ing.6 Nearly every vacuum column operating above a 730°F–740°F Fuel Steam
(388°C–393°C) flash-zone temperature has coked the wash section
packing in less than a four-year run.2 An inadequate wash-zone VRES
liquid rate is one of the primary causes for coking.7 The bottom
of the wash section is kept wetted by flash-zone entrainment. The FIG. 1 Flow diagram of a typical crude vacuum unit.
top of the packing is wetted by the wash oil flowrate.8
HYDROCARBON PROCESSING MARCH 2009 69
I
a vacuum unit, special attention should be paid to these critical ing the pressure profile accurately throughout the heater and
points. Vacuum unit product yields and critical operating condi- transfer line is important, because the heater-outlet and transfer-
tions must be accurately predicted.4 Features of the system are line pressures are used in the process model.4
the heater outlet, transfer line, flash zone, collector tray below the Estimating the heater-outlet and transfer-line pressure profiles
wash section and wash-section column internals.4 Other parts of accurately requires a model that is capable of rigorous tube-by-
the vacuum column are straightforward and well understood.4 tube heat transfer and accurate two-phase flow calculations.4
Often, the design of the wash section is considered a trivial Calculated phase regimes in the transfer line are either stratified
item; yet, process and equipment design issues surrounding the or stratified wavy.8,10 Stratified phases cause the liquid and vapor
wash section are complex.7 Wash-zone packing coking is caused to have poor mass and energy exchange across the interface.4,8
by poor feed characterization, process modeling and equipment Thus, liquid and vapor contact is poor.8 Since the transfer line
design.7 Wash-zone design and operation are not trivial issues.7 consists of large-diameter piping, the liquid and vapor separate
Predicting total VGO yield, operating temperature at the heater in the horizontal section of the transfer line, vapor flows along the
outlet and flash zone and wash-oil flowrate needed to prevent top of the pipe and liquid flows across the bottom.4,8 Transfer-line
coking are critical design parameters.4 Transfer-line, flash-zone vapor becomes superheated due to pressure reduction as the two
and wash-section designs influence the coking rate in the wash- phases approach the flash zone.4 Phase separation causes super-
section internals.10 heated vapor to flow through the top of the pipe and colder liquid
Vapor and liquid feed enter the column at velocities as high to flow on the bottom.10 Thus, the vapor and liquid entering the
as 380–400 ft/sec.4,6,8 The vapor phase contains small droplets of flash zone are not in equilibrium.4,8
VRES that have been generated in the transfer line. The droplet Assuming that the liquid and vapor entering a vacuum-column
size is too small to allow settling in the transfer line because the flash zone are in equilibrium is a critical mistake.4 Transfer-line
velocity is too high.4,6,8 Hence, the flash zone and wash sections phase separation increases the amount of wash-oil flow needed to
need to remove the entrainment.6 The flash-zone vapor horn and prevent coking, because the wash oil vaporizes more of the wash
flash zone help remove larger droplets and distribute the rising liquid.4 In reality, accounting for transfer-line phase separation
vapor across the column cross-section.6 By uniformly distributing raises the wash-oil flowrate by 200% to 300% over conventional
vapor, the high-velocity areas are minimized, allowing the packing modeling practices that assume liquid and vapor leaving the trans-
to remove essentially all of the small droplet residue.6 fer line are in equilibrium.8
In the vacuum unit, the transfer-line critical flow expansion, Often, the vacuum unit is modeled assuming that the liquid and
flash zone vapor horn and wash-section internals determine the vapor in the flash zone are in equilibrium.7 Assuming that the flash
amount of entrainment.2 The quantity of entrainment on a unit zone is in equilibrium, this position will cause the calculated wash-
varies according to the flash-zone design, flash-zone height, trans- oil rate to be too low.10 The vapor/liquid equilibrium may exist at
fer-line velocity, etc.9 Poorly designed transfer lines with high the heater tube outlet, but it does not exist in the flash zone.7
pressure drop critical flow expansions at the column inlet nozzle A practical approach to modeling transfer lines and vacuum
generate fine mists that are difficult to remove.2 Yet, the entrain- columns that better predicts yields and other critical operat-
ment can be almost eliminated through prudent transfer-line and ing parameters requires that the model to be segmented into a
column internal designs.2 number of operations before the vapor enters the column wash
While entrainment from the flash zone contains high metals, section.4 Using multiple unit operations allows estimating the
concarbon and asphaltenes, the amount of entrainment should non-equilibrium nature of the system.4,2
be minimized as much as possible. Transfer-line, flash-zone
and wash-section designs influence the HVGO concarbon, Evaluating different vacuum unit models. As men-
metals and asphaltenes content through their impact on Vac- tioned earlier, the sections that are important and critical that
uum residual (VRES) entrainment.10 The wash zone removes require to be accurately simulated are heater outlet, transfer line,
entrained residue from the flash-zone vapor and provides some flash zone and wash zone. Other parts of the vacuum column are
fractionation of the HVGO product.7,8 So, in the vacuum col- straightforward and well understood. While the entire unit will
umn design, flash-zone vapor entrainment and its effect on the be simulated, we will only use these listed sections to analyze and
wash zone should be considered, and the HVGO quality has evaluate different models. To evaluate different cases, simulation
to be calculated. Depending on the design, flash-zone vapor models were made according to these rules:
entrainment can enter the wash bed. Since the wash-section • Two theoretical stages were applied for the wash bed.
internals remove entrained VRES from the flash zone, liquid on • The heater outlet temperature was set for a TBP cut point of
the collector tray below the wash bed consists of true over-flash
plus removed entrainment from the flash zone.4 This liquid is TABLE 1. Simulation results of an ideal model
always referred to as slop wax. (equilibrium in the transfer line and no entrainment
to the wash zone)
Vacuum unit model. According to the mentioned criteria,
the critical sections of the vacuum unit are the fired heater, trans-
HVGO distillation
tail—95%-EP, °C
Minimum wash
zone minimum
HVGO 95%, °C
fer line, flash zone and wash section. Modeling other compo-
Place of wash
VRES 5%, °C
HVGO EP, °C
zone liquid
flow, m3/hr
nents of the unit are not complex and can be simply made and/or
rate, m3/hr
liquid rate
Wash-oil
predicted. When building a model to estimate critical operating

parameters, some simulation exercises are needed. But the problem
is: Can we believe the simulation results?
The only way to ensure that the model is representative of the 165 25 Bottom of 564 584 20 533
vacuum unit is to verify it against measured plant data.4 Estimat- wash zone
70
1,000°F (538°C) on the HVGO cut. The heater outlet was within in the middle of the wash zone. While the middle of the wash
the normal range for such a TBP cutpoint. section is prone to coking, it means that minimum liquid flow is
• All slop wax was sent to the top of the stripping section. occurring. Thus, simulation results that include entrainment in
• Flash-zone pressure, transfer-line pressure drop and, conse- the middle of the wash section are in complete agreement with the
quently, heater-outlet pressure were fixed for all cases. actual performance of the crude vacuum-tower wash section.
• The amount of entrainment from the flash zone is the same So, an estimated amount of entrainment should be considered
in all cases. in the simulation model. Table 2 shows the simulation results for
• The tower top pressure and temperature for all cases are the this case.
same. When compared against the ideal model, except for the
• The same amount of stripping steam was used for all cases. minimum wash-zone liquid flow, no considerable changes have
• The same number of theoretical stages was assumed on the occurred. In the equilibrium TL, entrainment from the flash
stripping section. zone has little effect on tower operating conditions and product
• A minimum wetting rate of 0.15 gpm/ft2 for the wash zone specifications for HVGO and VRES. The minimum wash-zone
was set on all cases. liquid for the ideal flash zone (no entrainment) is 25 m3/hr. This
At the first step, an ideal model is considered and simulated. In is true over flash. For the non-ideal flash zone (entrainment with
this ideal model, we will assume that the liquid and vapor phase the flash-zone vapor outlet), the minimum wash-zone liquid is 48
entering the tower flash zone are in equilibrium and that no phase m3/hr, which is not a true over flash. The entrained liquid droplets
separation occurs in the transfer line. Also, complete phase separa- from the FZ contain coke particles.
tion in the flash zone is considered (no entrainment). Table 1 lists When the droplets contact the wash-zone packing, coke parti-
the simulation results. cles transfer onto the packing surface. Liquid flow in the bottom of
Another case is an equilibrium transfer line (TL) with a non- the wash section is sufficient to remove the coke particles, and the
ideal flash zone (FZ) (considering an estimated amount of entrain- coke is transferred with the liquid. But, in the middle of the wash
ment). But the problem is how the entrainment could be entered section, conditions are different. Here, liquid flow is minimal.
into the simulation model. To answer this question, it is necessary If this flow is not sufficient, coke particles are not washed away.
to go through the process of what is happening in the vacuum- In such cases, the coke particles accumulate in the middle of the
tower flash zone. The vapor and liquid phases from the transfer wash section. By this view, the minimum wash liquid flow should
line enter the flash zone. Due to high velocity, a considerable be calculated based on the required liquid flow to remove and to
portion of the liquid is dispersed into the vapor phase as large
and small droplets. As mentioned earlier, the large droplets are TABLE 3. Simulation results of non-equilibrium TL with
removed by the flash-zone vapor horn and the flash zone. The no entrainment to the wash zone
wash zone removes small entrainment droplets from the flash-
zone vapor. Accordingly, the entrainment is the small droplets
HVGO distillation
tail—95%-EP, °C
Minimum wash
that are coming up with the flash-zone vapor.

zone minimum
HVGO 95%, °C
Place of wash
VRES 5%, °C
HVGO EP, °C
In the wash section, the small droplets are removed from
zone liquid
flow, m3/hr
rate, m3/hr
liquid rate
Wash-oil
the vapor phase. The removed droplets with the wash oil (over
flash), as a liquid phase, come down to the collector tray below
the wash zone. De-entrainment could happen in the middle of
the wash section. Thus, the entrained droplets could come up to 144 9 Bottom of 577 598 21 523
the middle of the wash-zone packing. In fact, from the bottom wash zone
to the middle of the wash-zone packing, the vapor phase from
the flash zone is in contact with the remaining wash oil, and the
separated droplets that are now coming down as a liquid phase
to the collector tray below the wash section. If the wash section Wash oil
is simulated by this viewpoint, the result should be proved with Wash
the reality of the vacuum tower. Transfer line zone
The simulation result of the tower, considering that the liquid vapor
Furnace Flash
entrainment comes up to the middle of the wash section, shows
outlet
that minimum wash-zone liquid flow happens just in the middle Transfer
Flash Flash
of the wash zone. As mentioned before, coke is always formed line
Overflash
Entrainment
TABLE 2. Simulation results of equilibrium TL with Transfer Flash
entrainment to the wash zone line liquid
HVGO distillation
tail—95%-EP, °C
Splitter
Minimum wash
zone minimum
HVGO 95%, °C
Place of wash
VRES 5%, °C
Stripping
HVGO EP, °C
zone liquid
flow, m3/hr
rate, m3/hr
liquid rate
Steam section
Wash-oil
VRES
167 48 Middle of 565 586 21 533 FIG. 2 Multiple unit operation for a non-equilibrium transfer line
wash zone model.

I
TABLE 5. Simulation results of non-equilibrium TL, non-

Wash oil ideal flash zone and no entrainment to the wash zone
Wash
HVGO distillation
tail—95%-EP, °C
Minimum wash
zone
zone minimum
HVGO 95%, °C
Transfer line
Place of wash
VRES 5%, °C
HVGO EP, °C
vapor
zone liquid
flow, m3/hr
rate, m3/hr
liquid rate
Furnace Flash
Wash-oil
outlet
Transfer
Flash Flash line
Overflash
137 9 Bottom of 577 599 22 521
Flash
wash zone
Entrainment
TABLE 6. Simulation results of non-equilibrium TL,

Splitter non-ideal flash zone with entrainment to the wash zone
Stripping
Steam section
HVGO distillation
tail—95%-EP, °C
Minimum wash
zone minimum
HVGO 95%, °C
Place of wash
VRES 5%, °C
HVGO EP, °C
VRES
zone liquid
flow, m3/hr
rate, m3/hr
liquid rate
Wash-oil
FIG. 3 Multiple unit operation for a non-equilibrium transfer line
with entrainment to the wash zone (modified model).
155 41 Middle of 569 591 22 527

TABLE 4. Simulation results of non-equilibrium TL with wash zone
entrainment to the wash zone (modified model)
a stripping section. The wash and pumparound sections of the
HVGO distillation
tail—95%-EP, °C
Minimum wash
zone minimum
vacuum column are modeled using a standard distillation col-

HVGO 95%, °C
Place of wash
VRES 5%, °C
HVGO EP, °C
umn model. The bottom-product stream from the distillation

zone liquid
flow, m3/hr
rate, m3/hr
liquid rate
Wash-oil
column is the true overflash. Entrainment and overflash feed an

adiabatic flash, with the operating pressure set at the pressure of
the collector tray located above the flash zone. Vapor feed to the
164 42 Middle of 568 591 23 529 wash section consists of transfer line vapor, collector tray vapor
wash zone and flash-zone vapor.
In this model, the maximum phase separation in the transfer
transport coke particles from the wash-bed packing surface and line has been considered. And, consequently, super-heated vapor
layers. This required liquid flow would be much higher than the enters the column. As seen in Fig. 2, entrainment was allowed,
minimum liquid flow to prevent the wash bed from drying out. but no contact between removed entrainment liquid and vapor
It is obvious that, the higher the FZ temperature, higher coke from the flash zone has been considered. Based on this proposed
particles will be produced. Actually, when the coke particle con- configuration, a simulation model was prepared and run. Table 3
tent of the entrained liquid droplet is increasing, the required summarizes the results from this simulation.
liquid for washing, removing and transporting the coke within the From Table 3, the results show, using this arrangement and with
wash-zone packing should be sufficient. If the liquid flow is not the same heater outlet, the wash-oil rate and minimum wash-zone
sufficient, then the coke particles can accumulate. Consequently, liquid flowrate were largely decreased. Also, the HVGO 95% and
the wash bed will coke up soon. For these conditions, nearly every EP increased. Conversely, a large drop in the VRES 5% occurred.
vacuum column operating above a 730°F–740°F (388°C–393°C) There are some discrepancies between the proposed arrange-
flash-zone temperature has lost wash-section packing due to coke ment and the real FZ (Fig. 1) configuration:
in less than a four-year run.2 1. By the recommended model, no contact between the liquid
A model has been proposed to address this non-equilibrium stream, which is produced from de-entrainment action of the
system.2,4 Fig. 2 shows a schematic of this model. In this model, wash zone, and vapor from the flash zone was considered.
vacuum unit operations consist of a simple exchanger (fired 2. Conversely, by using this model, the minimum wash-sec-
heater), with the outlet temperature determined by the HVGO tion liquid flow occurs in the bottom of the wash zone. In fact,
cutpoint target. The heater outlet pressure depends on the trans- this model could not predict coking of the middle of the wash-
fer-line pressure drop and whether parts of this line operate at zone packing.
critical two-phase velocity. 3. The transfer-line vapor and liquid with the stripper-section
The transfer line is modeled as an adiabatic flash, with the vapor outlet (strippout), are already in contact with each other in
pressure set at the same pressure as the first large horizontal sec- the real flash zone. As mentioned before, the vacuum tower flash
tion of the transfer line. Liquid and vapor from the transfer-line zone is not an ideal stage. So, the heat and mass transfer at this
flash are separated into two streams. The transfer-line liquid stage could not be done up to a theoretical stage (vapor and liquid
stream is split into an estimated flash-zone entrainment and outlet in equilibrium). But, in the proposed model, they meet
flash-zone liquid feed. each other at the theoretical stages.
The column flash zone is modeled as a simple flash if it does To correct the proposed model for discrepancies Nos. 1 and
not have a stripping section or as a distillation column if it has 2, modifications on the liquid entrainment could be considered.
72
The modified proposed model is shown in Fig. 3. Table 4 shows stage. A model is presented in Fig. 4 to solve this problem. In this
simulation results for the modified model. This simulation shows model, the phase separation and, consequently, super heating of
that, for the modified model, the minimum wash-section liquid vapor in the transfer line is considered. The vacuum tower is mod-
flow occurs in the middle of the wash zone. eled according to the standard simulation route.
Contrary to the equilibrium TL model, the effects of entrain- But, to compensate for non-idealities of the flash zone, a
ment on the operating conditions and HVGO specifications are non-equilibrium stage is determined. A model was developed to
considerable and are important for non-equilibrium TL models. simulate this case. The simulation was adapted to have the same
As seen, the entrainment to the middle of the wash section in amount of overflash to meet the specified minimum wetting rates.
the model causes the wash-oil rate, and minimum wash-zone Table 5 lists the simulation results for this case. The simulation
liquid flow increased from 144 m3/hr to 164 m3/hr and from results show some interesting points. In comparison to a similar
9 m3/hr to 42 m3/hr, respectively. The results also contain a model (the proposed model in Fig. 2), the lower wash-oil rate
considerable reduction in HVGO 95% and EP while the VRES was calculated as 144 m3/hr as compared to 137 m3/hr or the
5% increased. equivalent to 5.1%. The changes in the HVGO specifications and
All of the data express improvement in fractionation. In fact, VRES specs are not too much.
any contact of the superheated vapor from the flash zone with the In this model, entrainment from the FZ to the wash section
liquid from the de-entrainment action of the wash zone causes could be considered. In this case, a model will be made as shown
gains in fractionation. This is true because superheating of the
vapor phase in the transfer line occurs due to phase separation, TABLE 8. Simulation results for the case that all
which causes poor mass and energy exchange; thus, any con- non-idealities have summarized to the FZ stage with
tact between the vapor and liquid can lead to equilibrium. The entrainment to the wash section
maximal separation and fractionation are done when the transfer
HVGO distillation
tail—95%-EP, °C
line vapor and liquid are in equilibrium. In this case, there is
Minimum wash
zone minimum
HVGO 95%, °C
Place of wash
VRES 5%, °C
HVGO EP, °C
non-equilibrium TL, which produces super-heated vapor at the
zone liquid
flow, m3/hr
rate, m3/hr
liquid rate
column inlet.
Wash-oil
Unlike the expectation, the existing entrainment is useful in heat

and mass transfer point because it approaches the conditions (sys-
tems) to the equilibrium. But plugging of the wash-zone packing is 155 41 Middle of 569 591 22 527
very harmful and has caused unscheduled unit shutdown repeatedly wash zone
and/or periodically. Entrainment from the flash zone can plug off
the wash-section packing because it contains coke particles.
By modifying, two discrepancies were solved. Yet, there is one
Wash oil
more item to be resolved. This point is the non-ideal flash-zone
Transfer Entrainment
TABLE 7. Simulation results for the case that all line vapor
non-idealities have summarized to the FZ stage without Furnace
Non-ideal
entrainment to the wash section stage for FZ
outlet
Flash Flash Transfer
line
HVGO distillation
tail—95%-EP, °C
Minimum wash
zone minimum
HVGO 95%, °C
Place of wash
VRES 5%, °C
HVGO EP, °C
zone liquid
flow, m3/hr
rate, m3/hr
Splitter
liquid rate
Wash-oil
Steam VRES
FIG. 5 Flow diagram of a non-equilibrium transfer line, non-ideal

137 9 Bottom of 577 599 22 521
stage for the flash zone with entrainment to the wash
wash zone zone.
Transfer line vapor Wash oil Wash oil
Furnace
outlet
Non-ideal Heater
Flash Flash Transfer stage for FZ
Non-ideal
line stage for
TL and FZ
Steam Steam
VRES VRES
FIG. 4 Flow diagram of a non-equilibrium transfer line, non-ideal FIG. 6 Summarized conditions for a non-equilibrium transfer line
stage for the flash zone and no entrainment to the wash and a non-ideal flash zone in the non-ideal stage for the
zone. flash zone with no entrainment to the wash zone.

I
is one of the worst events in a vacuum unit and requires unit shut-
down to replace packing. So, although entrainment may push the
Wash oil system to higher yields or quality (in mass and energy exchange
Entrainment points of view), it can plug the wash section of the tower.
Heater
According to the presented study, under equilibrium for the
Non-ideal
stage for TL, no change will occur if entrainment is considered. When
TL and FZ the equilibrium TL provides vapor and liquid phase in the equi-
librium state and maximum mass and energy exchanges have
occurred, no more mass and heat transfer can be expected. So,
Steam while the desirable effect of entrainment could be achieved by
equilibrium transfer line, it is offered to eliminate the entrainment.
VRES New technology should address these goals:
• Provide equilibrium transfer line
FIG. 7 Summarized conditions for a non-equilibrium transfer line • Provide a suitable flash-zone arrangement and vapor horn to
and a non-ideal stage for the flash zone with entrainment eliminate entrainment from the flash-zone vapor outlet as much
to the wash zone.
as possible.
Currently, there are many designs for flash-zone arrangements
in Fig. 5. This model has all of the non-idealities for the transfer and vapor horns to eliminate entrainment. In some, the center
line and flash zone. The flash zone non-idealities consist of non- inlet is recommended; in others, a tangential type is offered. In
ideality in phase separation, and heat and mass transfer. It seems addition, the flash zones are available in different designs to remove
that the model (Fig. 5) could manage the realities found in crude entrainment from the flash-zone vapor outlet. Some designs are
vacuum towers. found in the open literature while the others are patented. Again,
Simulation results of this model are listed in Table 6. Again if the flash-zone arrangement is designed to remove entrainment
a noticeable change in the minimum wash-zone liquid flow without any attempt to maintain equilibrium in the transfer line,
occurred—137 m3/hr compared to 155 m3/hr or equivalent to then the quality and/or yield of the VGOs will drop.
13.1%. Also, decreases in HVGO 95% EP and increases in VRES
5% are considerable. Likewise, in the previous case, entrainment Options. When simulating crude vacuum units, some non-
to the middle of the wash section can compensate for many non- idealities must be considered. When developing a model based on
idealities in the TL and FZ and help the unit approach equilib- these non-idealities, these non-idealities must be identified and
rium to improve fractionation. This is obvious in simulation understood. The next step is to incorporate these non-idealities
results, as shown in Table 6. into the simulation model. While there are many options and
The question now is: Is it possible to summarize all non-ide- alternatives to develop simulation models, in some cases, a simple
alities of the TL and FZ in mass and heat transfer to the assumed model may be offered instead of sophisticated ones. As shown here,
non-ideal stage for the FZ? To answer this question, the model by a simple non-idealities assumption, a model was developed that
from Fig. 6 is considered. This model was simulated, and the is completely consistent to the real performance of the tower. HP
results listed in Table 7. This simulation was done to have the
same amount of overflash. The results are exactly similar to the LITERATURE CITED
1 Yahyaabadi, R., “Improve design strategies for refinery vacuum tower,”
case when phase separation is considered for the transfer line.
Hydrocarbon Processing, December 2007, p. 106.
For this case also, if entrainment from the FZ to the wash 2 Golden, S. W., T. Barletta, S. White, “Vacuum unit design for high metals
section is considered, a model as shown in Fig. 7 should be used; crudes,” Petroleum Technology Quarterly, Winter 2007, p. 31.
Table 8 lists simulation results for this case. The values from Table 3 Golden, S., “Canadian crude processing challenges,” Petroleum Technology
8 are exactly similar to a case in which the non-idealities were Quarterly, Winter 2008, p. 53.
4 Barletta, T. and S. W. Golden, “Deep-cut vacuum unit design,” Petroleum
addressed in the TL separately. Technology Quarterly, Autumn 2005, p. 91.
5 Golden, S. W. and T. Barletta, “Designing vacuum units,” Petroleum
What should technology do? As seen, considering the Technology Quarterly, Spring 2006, p. 105.
6 Golden, S. W., “Revamps: maximum asset utilisation,” Petroleum Technology
entrainment from the flash zone to the middle of the wash section,
Quarterly, Winter 2005, p. 37.
it corresponds with actual experiences from the crude vacuum 7 Golden, S. W., “Troubleshooting vacuum unit revamps,” Petroleum Technology
unit in many refineries. Furthermore, phase separation in the TL Quarterly, Summer 1998, p. 107.
and, consequently, creating superheated vapor at the tower inlet 8 Martin, G. R., “Vacuum unit design effect on operating variables,” Petroleum
has been discussed. According to the presented study, entrainment Technology Quarterly, Summer 2002, p. 85.
9 Golden, S. W., N. P. Lieberman and E. T. Lieberman, “Troubleshoot vacuum
from the FZ is not totally undesirable. In the non-equilibrium columns with low-capital methods,” Hydrocarbon Processing, July 1993, p. 81.
TL, the liquid and vapor phases do not have sufficient mass and 10 Hanson, D. and M. Martine, “Low capital revamp increases vacuum gas oil
energy exchange. In this case, the de-entrainment action of the yield,” Oil & Gas Journal, March 18, 2002.
wash section provides another opportunity for more mass and
heat exchange between the liquid and vapor phases from the
TL to approach equilibrium. Therefore, it is an improvement
because, in equilibrium, maximum mass and heat transfer occur.
Alternately, entrainment can plug the wash section due to coke Reza Yahyaabadi is a senior process engineer for Esfahan Oil
particles caused by cracking. Refining Co. (EORC), Esfahan, Iran. He has 20 years of experience
in process engineering, process revamps, debottlenecking and
Plugging the wash section causes low quality and yield of simulation and holds a BS degree in chemical engineering from
VGOs; all reduce plant profitability. Plugging of the wash section Esfahan University of Technology.
74
Showcase
u p s t re a m / m i d s t re a m / d o w n s t re a m
A Supplement to
&
The following companies are display advertisers in the Spring 2009 edition of the Upstream/Downstream Software Reference Guide. You can
access the entire guide online at www.gulfpub.com/gpc/. This edition will also be available at many key industry meetings such as GPA, NPRA
Spring Annual, SPE/IADC Drilling, SPE Production Ops, API Spring Annual, OTC, NPRA Maintenance, ILTA, AAPG and EAGE.
CD-adapco is the leading Gulf Publishing Company’s Software M3 Technoloy is

global provider of full-spec- Division publishes and distributes more a leading provider
trum engineering simula- than 40 desktop applications designed of ad-vanced asset
tion (CAE) solutions for fluid flow, heat trans- specifically for the needs of the engineering scheduling and optimization solutions for oil refin-
fer and stress, including Computational Fluid community involved in the petroleum industry. ing, petrochemical, natural gas-LNG and terminal
Dynamics (CFD/CAE) flow, thermal, and stress operating industries. M3’s SIMTO™ software cap-
software, CAD-embedded CFD options, and Haverly Systems Inc. is an inde- tures eco-nomic opportunities and reduces the cost
CAE consulting services and training. Its princi- pendent software company spe- of managing complex facilities at the plant level,
pal offices are in New York, London and Yoko- cializing in developing optimi- regional operating level and global enterprise level.
hama with subsidiary offices across the world. zation-related products and services for over four Select 415 at www.HydrocarbonProcessing.com/RS
decades. Their systems are used in over 50 countries
by international and independent oil companies as Paradigm™ is a lead-
Chemstations is a lead- well as petrochemical companies. The effectiveness ing provider of enter-
ing global supplier of proof their work has long been recognized in the con- prise software solutions for digital subsurface asset
cess simulation software tinued patronage and goodwill of our clients. management serving the global oil and natural gas
for the following process industries: oil & gas, Select 410 at www.HydrocarbonProcessing.com/RS exploration and production industry. Paradigm
petrochemicals, chemicals and fine chemicals, software solutions and integrated services for
including pharmaceuticals. We currently offer Heat Transfer Research Inc. prospect generation and evaluation support every
several individually licensed, and tightly inte- (HTRI) is an international stage in the oil and gas E&P process, from seismic
grated, technologies to address the needs of the consortium founded in 1962 processing and imaging, interpretation and mod-
chemical engineer, whether doing new process that conducts industrially relevant research and pro- eling, reservoir characterization and engineering,
design or working in the plant. vides software tools for design, rating and simula- well planning and drilling and data management.
tion of process heat transfer equipment. HTRI also Select 418 at www.HydrocarbonProcessing.com/RS
produces a wide range of technical publications and
Codeware is the leading provides other services including contract research, Schlumberger Infor-
provider of software for software development, consulting and training. mation Solutions offers
the design and analysis of Select 411 at www.HydrocarbonProcessing.com/RS a unique combination of E&P domain tech-
ASME Section VIII pressure vessels and heat nology, IT services and consulting experience,
exchangers. Since 1985, Codeware’s Engineers KBC Advanced Technologies is enabling oil and gas companies to discover bet-
have focused exclusively on meeting the engi- a leading independent consulting ter growth opportunities, drill better wells and
neering software needs of designers and users of and technology services group, de- improve production at lower cost.
vessels and exchangers. Over 1,000 companies livering improvements in business performance
currently rely on Codeware products. and asset value to owners and operators in the Soteica is a Process Engineering
oil refining, petrochemical and other processing Software Company dedicated to the
industries worldwide. KBC’s specialized soft- development, implementation, sup-
The Equity Engineer- ware improves profitability through industry- port and sustainability of its hydro-
ing Group, Inc. is a leading models and simulation technology. carbon and energy management applications, S-
recognized leader on ag- Select 412 at www.HydrocarbonProcessing.com/RS TMS and Visual MESA. For more than 20 years,
ing infrastructure fixed equipment service and Soteica has successfully deployed its applications
support for the oil and gas industry. Equity Merrick Systems in refineries and petrochemical plants throughout
helps plants manage risk and improve profit- provides integrated the Americas and Europe. Our customers have
ability with cutting-edge software and con- software solutions for achieved millions of dollars in recurring benefits
sulting strategies that maximize equipment oil and gas production operations. Since 1989, from the sustained use of our applications.
operational availability, control inspection Merrick’s product suite has grown to include Select 420 at www.HydrocarbonProcessing.com/RS
costs and avoid costly shutdowns. Best-of-Class applications for field data capture,
production hydrocarbon accounting, plant al- Spiral Software was set up in 1998
locations, regulatory reporting, production data in response to an observed change in
geoLOGIC systems ltd. pro- management and production monitoring. the needs of the evolving oil indus-
vides well data and integrated Select 416 at www.HydrocarbonProcessing.com/RS try. The modern oil industry faces a
software solutions to the energy growing challenge: while environ-
and production industry. The company is an m:pro IT Consult is a project ser- mental specifications on products grow tighter
innovator in supplying data in more accessible vices and software products com- each year, many of the remaining reservoirs offer
and usable forms so clients can make better de- pany which enables petroleum relatively poor quality crude oil. In this context,
cisions - from the well head to senior levels of refining, petrochemical and other industries to understanding the detailed refining behaviour
accounting and administration. achieve total integration of information sources of crude oil is more critical than ever before.
and applications, from business systems, ERP Spiral Software provides tools and services that
and supply chain management through to plant enable our clients to make the best possible
information, production planning, scheduling choices in trading and refining crude oil.
and operations decision support. Select 421 at www.HydrocarbonProcessing.com/RS
ARE YOU A
SUBSCRIBER?
Order online by April 15 to receive
these items in the coming months.
UBSCRIBE As a paid
S
/R
subscriber,
CLICK
in addition to your
ENEW
12 monthly issues, in print or digital:
• Online access to the current issue
and all the latest Process Handbooks
• Online access to the world’s most
powerful archive of HPI information
containing eight years of back issues
• Online subject/author index of print
articles with links to articles currently
available online.
• Monthly e-newsletters providing
an early preview of upcoming special
editorial features, which provide
operational and technical insights.
HYDROCARBON PROCESSING is
the leading monthly magazine for
staying connected to the hydrocarbon
processing industry. Published since 1922,
HYDROCARBON PROCESSING provides
operational and technical information to
improve plant reliability, profitability, safety
and end-product quality. The editors of
HYDROCARBON PROCESSING bring you first-
hand knowledge on the latest advances in
technologies and technical articles to help
you do your job more effectively.
Receive Monthly e-newsletters providing

an early preview of upcoming special
editorial features, which provide
operational and technical insights:
April 2009:
• Process licensors
• Catalyst developments
• Distillation methodologies –
towers, trays, packings
• Reactors, vessels and internals
2 simple ways
to subscribe:
• Visit www.HydrocarbonProcessing.com
• Call +1 (713) 520-4440
May 2009:
• Pumps, valves compressors and
driver reliability
• Corrosion
• Measuring and testing
• Heat Exchangers
June 2009:
• Process control technology,
real-time optimization
• Distillation/trays/packing
• Product yields via new technology
• Energy management – furnaces, boilers,
heat exchangers
From dynamic ‘mysterious’ control

to dynamic ‘manageable’ control
Instructional design strategies and delivery methods for bridging
the DMC chasm
S. M. RANADE and E. TORRES, RWD Technologies LLC, Houston, Texas
D
ynamic matrix control (DMC) is a type of model-based ter estimation of tuning parameters, newer techniques to deal with
process control (MPC) that uses an explicit plant model integrating systems, integration with neural networks, etc. These
derived from plant tests. It has been over 25 years since areas of research are clearly needed to maintain their competitive
information about the first industrial applications of DMC edge but have not helped address the discomfort of newcomers
appeared in the open literature. Since then, hundreds of papers to the technology.
and many books have been published on the topic of MPC or • Traditional DMC courses have focused on training the site
model predictive control (MPC). specialists and process control engineers on how to use the soft-
Every new edition of all process control books typically has ware but not on the operators who run the unit. The courses in
a chapter devoted to MPC. DMC the past have focused more on prod-
experts and some plant process con- ■ "Initially, the operators need to uct features than on learners’ needs.
trol engineers and operators involved • Since its inception over 25
in the entire implementation cycle know the what and, to some extent, years ago, a new generation of opera-
of a DMC project seem to be com- tors and engineers has started work-
fortable with the technology. How-
the why. Understanding the how ing at the sites where the technology
ever, typical responses we get from will then happen with time." was implemented. At many sites, the
board operators (who run the units) specialists are gone and knowledge
include: —Ricardo Lecompte P., about DMC implementations has
• DMC probably makes money operations supervisor, been orphaned.
for the unit, but we are not sure how.
• After a recent thunderstorm, Ecopetrol, Cartagena, Colombia Proposed solution. Michael
the controller went crazy! Buckland and Doris Florian in their
• Sometimes I expect it to increase the flow, yet it seems to paper on intelligent information systems identified four courses
raise the pressure. I am not sure why. of actions when the complexity of a task strains or exceeds one’s
expertise: education, advice, simplification and delegation.2
The gap. With such a long history and ubiquitous presence in MPC software vendors are actively engaged in investigating
process plants, one would have expected a higher level of knowl- ways to improve user interfaces.3 Many universities have created
edge and acceptance of dynamic matrix controllers among opera- process control labs to improve their students’ familiarity with
tors and process engineers. This does not seem to be the case. the technology.4 We decided to approach the problem from a
Possible reasons for this gap include: training perspective.
• The technology is complex. This was recognized and dis-
cussed by Cutler, et al., in the early 80s.1 Even today, some “die- Different perspective. Instead of comparing monitoring a
hard” purists sense a high degree of risk in attempting to try to dynamic matrix controller to performing a brain surgery which
get everyone on board. Their advice is to let the experts handle made sense more than 25 years ago when Dr. Cutler and other
the problems. pioneers5 in the field were applying it in plants with limited
• In a typical DMC project, 80% or more of the investment computing power and high perceived risks, we began by asking
goes into the initial model identification, tuning and commis- a different question: “What if running a DMC application was
sioning part of the project. The assumption is that if designed and more like driving a computer-controlled car?”
tuned correctly the controller should run relatively maintenance- For operating and maintaining such a vehicle, one does not
free for a long time. While this is a reasonable assumption, it does have to be an expert on the design and tuning of the computer or
not mitigate the anxiety among the operators who run the units. the engine. Yet, having a foundational understanding of how the
• Recognizing the maturity of the core technology, MPC computer and car work together would clearly be beneficial to
software vendors have shifted their emphasis from knowledge everyone involved. This different perspective seemed to resonate
transfer to enhancements such as using state-space modeling, bet- well with users and lead to developing a DMC course.
I 77
Target audience, objectives. Using the terminology intro- instructional design is to ask: “Who will take this course?” We
duced by Guy Boy,6 we selected two goals for the course: increase selected the board operators and new plant engineers as our initial
user knowledge to improve cognitive stability and use simplifica- target audience and set the following objectives for the course:
tion to reduce perceived complexity. The most important step in On course completion, the students will be able to:
1. Communicate more precisely about DMC
2. Show a measurable improvement in their ability to monitor
Steady state
Marginal costs
their unit with DMC on it
Objective Gain matrix
3. Show an improvement in their ability to diagnose a prob-
LP
optimization Equal concern error lem in their DMC-controlled unit.
Priority group
module
Constraints
With a known learner-profile and defined objectives, we broke
LP step
down the content into units and selected a sequence and delivery
LP targets Time to steady state
Model horizon
methods. Two instructional design strategies and two instructional
Plant test Prediction horizon delivery methods got high ratings during the initial test runs of the
DMC:
operator Prediction model Control horizon course. The two instructional design strategies are:
view Unit response curves • Simplify content to match learner needs
Equal concern • Equip students with alternate schemas to validate new
Controller Objective Move suppression
knowledge.
Weighting/ranking
Move
The two instructional delivery methods are:
LP step
calculation
Max move Instrument limits
■
Use humor and metaphors
Constraints ■
Use dynamic interactive motion graphics elements.
Upper and lower limits Safety limits
Operator limits These strategies and methods can be easily extended to improve
Critical variables
the training effectiveness for other DMC-like advanced technolo-
gies and are the main focus of the rest of this article.
FIG. 1 Example of key DMC concepts from the literature.8–13
Simplify content. A literature review reveals that most
articles on MPC are written for application developers and
include details such as model horizon, control horizon, predic-
Handles t'JMMJOUIFCMBOL tion horizon, move suppression factors, coincidence points,
or MVs tMVs@@CVs. etc., that are of minimal interest to operators and new process
t6TFUIFFYUSBEFHSFFT engineers trying to learn the technology. However, simplifica-
0CKFDUJWF PGGSFFEPNUP
Obligations tion is a double-edged sword as illustrated by the well-known
IBWFGVO or CVs maxim called Ockham’s razor.7 Cognizant of the requirement
that any simplified representation of DMC should not preclude
future in-depth understanding of the application, we created a
“simplified” overview derived from numerous textbooks and
papers8–13 and shown in Appendix A: How DMC works. The
common terms used in association with DMC applications
are defined in Appendix B: Glossary. Fig. 1 is an example of a
system representation appropriate for our situation. The main
FIG. 2 Representation of the fictional Beauford’s life in his early concepts that must be understood by the learner are marked
20s. with the “key” icon.
Use humor and metaphors. The operators we interviewed

MVs ranked CV priority
felt challenged by the language and terminology used in connec-
by cost High Low tion with DMC applications. As reported by Benedict Carey in
The New York Times, researchers have found that the human brain
-10 LDL has a natural affinity for narrative construction.14 Also, emotional
Objective: memory has been recognized as the most effective pathway to
-5 retire peacefully
long-term retention.15
Author Marc Prensky coined the phrase “digital natives” to
represent the new generation of high school and college gradu-
-2 ates.16 According to Prensky, this new generation prefers games to
Angry ex “serious” work. So, instead of simply listing the DMC vocabulary,
we decided to leverage the power of stories, metaphors and humor.
We created a story called “The life of Beauford—A DMC inter-
BP pretation.” Figs. 2 and 3 illustrate snapshots of Beauford’s life in
5 DV: Ex +
lawyer
his early 20s and in his mid-50s.
For example, to illustrate the concept of equal concern error,
we stated that one phone call from his ex-wife’s lawyer, a 5-mpg
FIG. 3 Representation of the fictional Beauford’s life in his early drop in his car’s mileage, a 10-point drop in his son’s grade and
50s. a $500 drop in his bank account were all viewed by Beauford
78
■ "Humor and use of interactive models TABLE 1. Data for LP example

Prod. time per Prod. time per Max. time available
in this course helped me overcome an Plant light engine heavy-duty engine per week
internal barrier to learning." Abilene 1 0 4
Birmingham 0 2 12
—Maria Helena Calvachi, sr. process Calgary 3 2 18
engineer, Ecopetrol, Barranca, Colombia Profit per engine $300,000 $500,000
as deviations worthy of equal concern. The learners who saw concepts like constraints, feasible region, etc., and to demonstrate
Beauford’s story were able to quickly grasp and recall in their own that the maximum value of the objective function always occurs at
words concepts such objective function, manipulated variable a vertex of the graph. Since operator adaptation to DMC requires
(MV ), controlled variable (CV ), disturbance variable (DV ) and a shift from sensor-motoric mode to a cognitive mode,6 we sup-
even specific terms such as CV ranks, equal concern error and plemented the traditional approach with an interactive approach
dynamic equal concern error. to allow the learners to experience the “what if scenarios:” What
if the plant availability of the Birmingham plant decreased by an
Use dynamic interactive motion graphics. Digital hour per week? What if the profitability of each heavy-duty engine
natives, who will soon be joining the process industry, grew up decreased by $100,000?
playing video games and learn by interacting with the content. In the initial tests, the students were quickly able to discover
DMC has the right level of complexity to explore the use of and explain that their changing of the plant availability had an
“motion graphics” in training. The availability of simulation soft- immediate impact on the size and shape of the feasible region for
ware,17 animation programs and other graphics tools has also made optimization and that although they changed the conditions the
it easier to incorporate motion graphics in teaching. Here’s an optimum still occurred at a vertex of the graph. You may try this
example that illustrates the role and importance of animation. yourself by going to: http://elearning.rwd.com/dmc.
Consider the problem of optimizing the number of light- and This “interaction” approach to be ideal for introducing “time-
heavy-duty engines18 manufactured by an aircraft manufacturer dependent” concepts such as dead time and inverse response and
using facilities in Abilene (A), Texas; Birmingham (B), Alabama; to enable learners to experience how narrowing down the range
and Calgary (C), Canada. The light engine is made with parts of reflux or steam rates in a distillation column would constrict
produced in plants A and B. The heavy-duty engine is made with the movement of a dynamic matrix controller.
parts produced in plants B and C. As shown in Table 1, there are
constraints on the availability of the three facilities. The objec- Equip students with alternate schemas. Many years
tive is to find the mix of light- and heavy-duty engines that will ago, when I was a graduate student, I had calculated a heat trans-
maximize the weekly profits. fer coefficient for a heat exchanger. I shared the result with my
Traditionally, linear programming (LP) courses use static para- advisor, Dr. Prengle. He did a quick “back of the envelope” cal-
metric graphs such as the one shown in Fig. 4 for introducing the culation and told me that I might have missed a decimal point.
He was correct.
For true learning to occur, the learner has to process, cross-
Engine sales optimization check and validate new knowledge by some alternate means. Some
Constraint: Abilene, Birmingham and Calgary
profit contours of this knowledge validation happened, as in the case of the late
10 Dr. Prengle, through years of experience. When faced with a new
y≤4 technical problem, my natural instinct is to use the language of
9
x≤6 mathematics to analyze the situation. Operators and engineers
8 3x + 2y ≤ 18
Profit of $ 1,200,000
No. of light engines per week
Profit of $ 2,700,000
7 Profit of $ 3,600,000
Parameter Change Change DMC response
6
Operator-entered limits Feasible region
5 CV upper limit = CV lower Feasible region; CV at
limit setpoint
4 No. MVs = no. of CVs Unique solution
No. MVs relative to no. CVs Scope for economic
3 optimization
No. MVs relative to no. CVs Give up on some CV
2 limits by priority
ECE for CV limit (tolerance) Importance of that limit
1 LP cost Use of that MV (resource)
No. MVs constrained Feasible region; give up
0 on some CVs.
0 1 2 3 4 5 6 7 8 09 1 No. of available MVs Number of CVs @ limits
No. of heavy-duty engines per week
FIG. 4 Parametric plot to illustrate LP concepts. FIG. 5 New DMC tool: parameter relationship diagram and
description.

I 79
TABLE 2. What if analysis strategy example

Unit (Fig. 6) 1. Meet overhead product specs.
objectives: 2. Meet bottom product specs.
M 3. Minimize energy consumption.
40 Y For the unit
F F
Bottoms CV st ppm heavies ⌬P as proxy ppm lights
in OVHD for flooding in bottoms
FF FF CV Overhead
T1 T2 F 300 ppm LP costs MV s
A product
Feed CV 10 Steam flow SP
20
DP 0.75 bar
Bottoms 5 Reflux flow SP
cooler DV s
Steam Feed rate FI
1
CV priorities
F
M Feed Group 2 1 2
V F Bottom Upper limit High High Medium
product
MV Lower limit Low Low Low
CV Bottoms
2,500 ppm A Question: What if the flooding constraint becomes active?
cooler
Analysis: • Under normal operating conditions, the DMC controller has 2 CV s
FIG. 6 A distillation tower example originally presented by and 2 MV s
Hokanson, D.A. and J. G. Gerstle.13 • From Fig. 5, No. MV s = No. CV s t Unique solution. Also, No.
of available MV s = No. of CV s at constraint. i.e., the product
specs are being met.
new to DMC do not have the benefit of vast experience or other
• The flooding constraint active t loss of a degree of freedom.
internal reference or a mechanism to crosscheck and validate the
actions of DMC. • The controller has to give up on one of the two product specs.
To move their learning about DMC from Bloom’s recall level • From Fig. 5, CV priority t in feasible region.
to analysis level, they need some “back of the envelope” type • The controller will give up on the ppm lights in bottoms
tools. Of course, if one could easily do “quick” calculations to constraint but maintain OVHD purity.
get the same answer as that found by DMC, one would not • Since the LP cost factor for steam is higher than that for reflux,
need DMC. The true benefits of technologies like DMC are the controller will first reduce steam and then cut reflux to
in their power to find “non-intuitive” but optimum solutions. meet the LP targets.
One approach that might accelerate this “knowledge internal-
ization” process is to provide “learner-appropriate” tools that also indebted to the following individuals who without hesitation and with great
would enable students to make at least qualitative sense of the patience shared their knowledge about MPC and DMC: Professor Larry Ricker
actions taken by the controller. Such an approach would move of U. of Washington, Professor Michael Nikolaou of U. of Houston, Professor
the student from simply the “know what” to the “know why” Stephanie Guerlain of U. of Virginia, Javier Sanchis of Universidad Politecnica de
Valencia, Spain, and David H. Jones of KBR.
state of learning.
Our understanding of the learning styles and preferred infor-
LITERATURE CITED
mation processing methods of operators and new plant engi- 1 Cutler, C., et al., “An Industrial Perspective on Advanced Control,” AIChE
neers matches closely with the empirical theory of intelligence Annual Meeting, Washington, D.C., October 1983.
developed by Elliott Jacques.19 Recognizing that most new board 2 Buckland, M. K., and D. Florian, “Expertise, task complexity, and the
operators are comfortable with a “symbolic verbal” (vs. abstract role of intelligent information systems,” Journal of the American Society for
conceptual) style of learning and have experience in declarative, Information Science, 42(9), pp. 635-643, October 1991.
3 Guerlain, S., et al., “The MPC Elucidator: A case study in the design for
cumulative and serial methods of information processing, we human-automation interaction,” IEEE Transactions on Systems, Man, and
developed two “cognitive” tools and a recipe-type “what if analysis Cybernetics - Part A: Systems and Humans, 32(1), 25–40, 2002.
4 Cooper, D. J., and D. Dougherty, “Enhancing Process Control Education
strategy” that uses the two tools. One of these tools—parameter
relationship diagram and description—for DMC is illustrated with the Control Station Training Simulator,” Computer Applications in
Engineering Education, 7, 203, 1999.
in Fig. 5. An example of the application of the two tools and the 5 Cutler, C. R., “Dynamic matrix control—a computer control algorithm,”
“what if analysis strategy” to a refinery distillation column (Fig. AIChE National Meeting, Houston, Texas, April 1979.
6 Boy, G. A., “Perceived Complexity and Cognitive Stability in Human
6) is presented in Table 2.
These instructional design strategies and delivery methods Centered Design,” Proceedings of the HCI International Conference, Beijing,
China, 2007.
increase user knowledge and decrease the perceived complexity 7 A maxim attributed to William of Ockham—a 13th century English
of DMC. The techniques can be easily extended to designing Franciscan scholar: the fewest possible assumptions should be made in
and delivering training for other DMC-like mature complex explaining a thing.
8 Marlin, T. E., Process Control—Designing Processes and Control Systems for
applications. HP
Dynamic Performance, 2nd Edition, McGraw-Hill, Singapore, 2000.
9 Seborg, D. E, et al., Process Dynamics and Control, 2nd Ed., John Wiley and
ACKNOWLEDGMENT Sons, Inc., Hoboken, New Jersey, 2004.
10 Qin, S. J., and T. A. Badgwell, “A survey of industrial model predictive tech-
The authors would like to thank RWD LLC for funding this work. E-mail
exchanges with Professor James Riggs of Texas Tech University and discussions nology,” Control Engineering Practice, 11, 2003, pp. 733–764.
11 Sorensen, R. C., and C. R. Cutler, “LP integrates economics into dynamic
with George Dzyacky and George Ho-Tung of BP were beneficial. The authors are
80
3. The move calculation module then uses the setpoint targets, data from the
prediction model and disturbance variable data and finds the values of the MV
Valve position
Plant moves by minimizing weighted sum of squares of the deviations of controlled
variables from their setpoints. Factors such as move suppression and equal con-
cern error affect the performance of the dynamic controller but are typically set
Measured MV during the design and tuning phases of implementation. The DMC controller
data Regulatory setpoint
calculates several control moves depending on the set control horizon (another
controller
design tuning parameter) but only implements the first move. Once the first move
Unit step is implemented, the next control cycle begins.
Unmeasured
response DV DV
Plant data APPENDIX B: GLOSSARY 8–13
Move
step Prediction
calculation Term Definition
test
MV values CV Controlled variable. Similar to output variable. Usually
CV values, measured or monitored. May also be inferred. Setpoints and
and SS gains upper and lower limits are typically associated with a CV.
Steady state CV rank or priority This parameter is used by the controller to prioritize the
optimization Targets for CVs different CVs. It is especially relevant for determining a fea-
CV priority, ECE within and MVs sible region for optimization. The CVs are typically grouped
each priority group, MV depending on their importance.
costs, operator limits DMC Dynamic matrix control (originally developed by C.R.
on MVs Cutler and others (1979)
DV Disturbance variable. Sometimes referred to as a FFV or
FIG. A-1 A DMC-based controller has three main modules (blue). feed forward variable. There are two types of DVs: modeled
and not modeled. A typical example of an unmeasured DV
is ambient temperature.
matrix control,” Hydrocarbon Processing, p. 57, September 1998. ECE Equal concern error. It is of two types. ECE used in the
12 Emoto, G., et al., “Integrated Advanced Control and Closed-Loop Real-Time steady state optimization part is based on the operator con-
Optimization of An Olefins Plant,” IFAC, Advanced Control of Chemical cern for deviation of each CV from its limit. It is used to
Processes, Kyoto, Japan, 1994, p. 95. normalize the differences caused by different engineering
13 Hokanson, D. A. and J. G. Gerstle, “Dynamic Matrix Control Multivariable units. Smaller value of ECE implies higher importance of
Controllers,” Chapter 12 in Practical Distillation Control, Edited by Luyben, the setpoint of a CV. Dynamic ECE allows the controller to
W. L., Van Nostrand Reinhold, New York, 1992. take different actions depending on how far the prediction
14 Carey, B., “This is your life (and How you tell it),” The New York Times
is from the setpoint.
Health Section, May 22, 2007. MV Manipulated variable. Similar to input variable. Must be
15 Wyman, P., “High performance memory,” in www.howtolearn.com, Jan 23,
independent, i.e., must not depend on another manipulated
2006. variable. May be a PV or an OP. Examples include reflux
16 Prensky, M. “Digital Natives, Digital Immigrants,” from On the Horizon,
rate, steam valve position, etc.
(MCB University Press, 9(5), October 2001. LP cost It is derived from plant operating costs and is consistent with
17 Model Predictive Control Toolbox Version 2.3.1 for use with MATLAB, The
the operating objectives of each unit. This parameter is used
MathWorks, Natick, Massachusetts, 2007. to drive the MVs toward the economic optimum.
18 Feron, E., et al., in “Introduction to Linear Programming,” Course 16.410,
Operator set limits For each MV, the upper and lower limit that can be set and
MIT, Fall 2003 in web.mit.edu/16.410.
19 Elliott Jacques’ theory summarized by Howard, P. J., The Owner’s Manual for changed by the unit operator.
the Brain, 3rd Ed., p. 787, Bard Press, Austin, Texas, 2006. SSG or SS gain Steady state gain is the net change in the CV value after the
effects of the step change in the MV have settled out over
time. A measure of how much the CV is likely to change
APPENDIX A: HOW DMC WORKS8–13 relative to a change in the MV value. It is a coefficient used
In DMC, the dynamic matrix is generated from the plant step tests. The in the LP cost calculations.
identification process begins with understanding the unit objectives and selection
of the MVs, CVs and DVs. The step tests are conducted to capture data (numerical
and graphical) regarding how each controlled variable responds to a step change
in each manipulated variable. The unit step response curves are then used as the
prediction model.
As the overview in Fig A-1 illustrates, a DMC-based controller has three main Saidas “Sai” Ranade is the manager of process and product
modules: the prediction module, the steady state (SS) optimization module and innovation for RWD Technologies LLC in Houston, Texas. Dr. Ranade
the move calculation module. earned his PhD in chemical engineering from the University of
1. Each controller cycle begins with collection of measured or actual values Houston. Early in his career, he worked as a process engineer. He
of the controlled variables. A comparison is made between the actual and the pre- has extensive consulting experience in the fields of pinch technol-
dicted values (from the model), and the error is then fed to the prediction model. ogy, process simulation, process design, business process mapping and business
This error term is assumed to be the same for the prediction horizon used by the strategy development. Dr. Ranade is the winner of Ed McMahon’s Next Big Star
model and accounts for model mismatch with the plant. The Prediction module comedy competition and has also taught high school algebra in the Spring Branch
generates an estimate of the steady state values of the controlled variables. The Independent School District in Texas.
current values of the manipulated variables and the future steady state values of
the controlled variables are passed on to the SS optimization module.
2. The SS optimization module uses the controlled variable priority informa-
tion and the steady state gain information from the unit response curves to check Enrique Torres is a senior training engineer in RWD Tecnologies LLC, Colombia
the feasibility of finding a solution. Assuming that one or more feasible solutions office. He holds an MS degree in chemical engineering from New Mexico State
are found, the optimizer then uses the LP cost data to determine the economic University. Before joining RWD, Mr. Torres worked for 23 years for Ecopetrol S.A.
optimum and send these targets as “desirable” setpoints to the move calculation He started his career as a process engineer at the Cartagena refinery. He also held
module. The priority group structure of controlled variables enables DMC to find positions as project engineer and as logistics coordinator. Mr. Torres devoted the past
feasible solutions. DMC starts with the highest priority rank group of controlled seven years of his career at Ecopetrol to lead the automation and control group at the
variables. If no feasible solution is found, DMC uses the equal concern error R&D center. As a process control engineer, he has extensive experience in the areas of
(ECE) data within each rank group to find feasible solutions. APC, alarm rationalization and operator training systems development.

I 81
SOFTWARE AND INSTRUMENTATION HPI MARKETPLACE
CA
Co PE-O
mp PE
lian N
t!
HTRI Xchanger Suite® – an integrated, easy-to-use suite of tools that

delivers accurate design calculations for
• shell-and-tube heat exchangers • fired heaters
• jacketed-pipe heat exchangers • air coolers
• hairpin heat exchangers • economizers
• plate-and-frame heat exchangers • tube layouts
• spiral plate heat exchangers • vibration analysis
Interfaces with many process simulator and physical property
packages either directly or via CAPE-OPEN.
Heat Transfer Research, Inc. [email protected]

150 Venture Drive www.HTRI.net
College Station, Texas 77845, USA
BUSINESS AND TECHNICAL SERVICES Select 209 at www.HydrocarbonProcessing.com/RS
0IPE3TRESS
0ROCESS3IMULATION
0ELLETIZING$IE$ESIGN
(EAT4RANSFER!NALYSIS
&INITE%LEMENT!NALYSIS
#OMPUTATIONAL&LUID$YNAMICS
6ESSEL%XCHANGER-ACHINE$ESIGN
2OTOR$YNAMICS3TRUCTURAL$YNAMICS
3PECIALISTSINDESIGNFAILURE Select 211 at www.HydrocarbonProcessing.com/RS

ANALYSISANDTROUBLESHOOTINGOF (OUSTON4EXAS
STATICANDROTATINGEQUIPMENT
4ELss
WWWKNIGHTHAWKCOM &AXss
CALL 713-520-4449
to advertise here!
NOW HIRING Select 212 at www.HydrocarbonProcessing.com/RS
WANTED:
MANUFACTURER’S
REPS
Dorf Ketal Chemicals seeks
Manufacturer’s Reps to represent
our process control chemicals Call 713/520-4449
in the Refinery market and the for details about
Hydrocarbon Processing’s
Ethylene sector. Ideal candidates
will have significant experience and Recruitment
live near such plants. Contact: Advertising Program
[email protected]
Use a combination of print, recruitment e-newsletter, plus Website to reach our
Select 213 at www.HydrocarbonProcessing.com/RS total audience circulation of more than 100,000 !

I 83
SALES OFFICES—EUROPE SALES OFFICES—OTHER AREAS
FRANCE, SPAIN, PORTUGAL, SOUTHERN AUSTRALIA – Perth
BELGIUM, LUXEMBOURG, SWITZERLAND, Brian Arnold
GERMANY, AUSTRIA, TURKEY Phone: +61 (8) 9332-9839, Fax: +61 (8) 9313-6442
2 Greenway Plaza, Suite 1020
Houston, Texas, 77046 USA Catherine Watkins E-mail: [email protected]
Houston, Texas 77252-2608 USA Ohana, 30 rue Paul Vaillant Couturier BRAZIL – São Paulo
Phone: +1 (713) 529-4301, Fax: +1 (713) 520-4433 78114 Magny-les-Hameaux, France Alfred Bilyk
www.HydrocarbonProcessing.com Tél.: +33 (0)1 30 47 92 51, Fax: +33 (0)1 30 47 92 40 Brazmedia Rua General Jardim, 633 Cj 61 01223 011
Mark Peters, Publisher E-mail: [email protected] São Paulo SP, Brazil
SALES OFFICES—NORTH AMERICA ITALY, EASTERN EUROPE Phone: +55 (11) 3237-3269
Fabio Potestá Fax: +55 (11) 3237-3269
AK, AR, AZ, CA, CO, ID, KS, LA, MO, MT, NM, NV,
Mediapoint & Communications SRL E-mail: [email protected]
OK, OR, TX, UT, WA, WY, WESTERN CANADA
Laura Kane Corte Lambruschini - Corso Buenos Aires, 8 JAPAN – Tokyo
2 Greenway Plaza, Suite 1020, Houston, Texas, 77046 5° Piano - Interno 7 Yoshinori Ikeda
Phone: +1 (713) 520-4449, Fax: +1 (713) 520-4459 16129 Genova - Italy Pacific Business Inc.
E-mail: [email protected] Phone: +39 (010) 570-4948, Fax: +39 (010) 553-0088 Phone: +81 (3) 3661-6138, Fax: +81 (3) 3661-6139
E-mail: [email protected] E-mail: [email protected]
AR, KS, LA, MO, OK, TX
Josh Mayer RUSSIA/FSU INDONESIA, MALAYSIA, SINGAPORE, THAILAND
5930 Royal Lane, Suite 201, Dallas, TX 75230 Lilia Fedotova Peggy Thay
Phone: +1 (972) 816-6745, Fax: +1 (972) 767-4442 Anik International & Co. Ltd. Publicitas Major Media (S) Pte Ltd
E-mail: [email protected] 10/2 Build. 1,B. Kharitonyevskii Lane Phone: +65 6836-2272, Fax: +65 6297-7302
CT, MA, ME, NH, NJ, NY, PA, RI, VT, 103062 Moscow, Russia E-mail: [email protected]
Phone: +7 (495) 628-10-333
EASTERN CANADA KOREA – Seoul
Merrie Lynch E-mail: [email protected]
Joong Hyon Kwon & JES MEDIA, INC>
20 Park Plaza, Suite 517, Boston, MA 02116 UNITED KINGDOM/SCANDINAVIA, NORTHERN Phone: +82 (2) 481-3411, FAX: +82 (2) 481-3414
Phone: +1 (617) 357-8190, Fax: +1 (617) 357-8194 E-mail: [email protected]
BELGIUM, THE NETHERLANDS
Roger Kingswell
PAKISTAN – Karachi
IA, IL, IN, KY, MI, MN, ND, NE, OH, SD, WI Gulf Publishing Co., P. O. Box 437 S. E. Ahmed
Janis Mason Maidstone, Kent Intermedia Communications
3711 Ravenswood Ave., Unit 146, Chicago, IL 60613 ME14 4RB United Kingdom Karachi-74700, Pakistan
Phone: +1 (773) 325-1804, Fax +1 (773) 325-9406 Phone: +44 (1622) 721222, Fax: +44 (1622) 721333 Phone: +92 (21) 663-4795, Fax: +92 (21) 663-4795
E-mail: [email protected] E-mail: [email protected]
AL, DC, DE, FL, GA, MD, MS, NC, SC, TN, VA, WV
Lee Nichols
CLASSIFIED ADVERTISING AND REPRINTS
2 Greenway Plaza, Suite 1020, Houston, Texas, 77046
Phone: +1 (713) 525-4626, Fax: +1 (713) 520-4433 Classified e-mail: [email protected]
E-mail: [email protected] Reprints e-mail: [email protected]
WinHeat 4 ©
Process Tools™
U d dF
Updated Features!!
Winheat suite provides all the tools for: For the operator in the petroleum and chemical industry, Process
Modeling most fired heater Tools opens a new door to user-friendly process calculations for the
Generating property grids design and evaluation of process equipment.
Insulation & heat loss computations
Features include:
Air preheat system analysis
Flare system design
Ducting design
Air cooled exchanger design
ID and FD fan sizing
Emission calculations from storage and/or vents
Tube wall thickness calculations
Heater draft analysis Order now!
Combustion analysis www.GulfPub.com/ProcessTools
Order now!
www.GulfPub.com/WinHeat Gulf Publishing Company
+1-713-520-4428 l +1-800-231-6275 l Email: [email protected]

84
FREE Product and Service Information — MARCH 2009
HOW TO USE THE INDEX: The FIRST NUMBER after the company name is the page on which an This information must be pro-
advertisement appears. The SECOND NUMBER, appearing in parentheses, after the company name, vided to process your request:
is the READER SERVICE NUMBER. There are several ways readers can obtain information: PRIMARY DIVISION OF INDUSTRY
1. The quickest way to request information from an advertiser or about an editorial item is to go to www. (check one only):
HydrocarbonProcessing.com/RS. If you follow the instructions on the screen your request will be forwarded for A 䊐-Refining Company
immediate action. B 䊐-Petrochemical Co.
C 䊐-Gas Processing Co.
2. Go online to the advertiser's Website listed below. F 䊐-Equipment Manufacturer
G 䊐-Supply Company
3. Circle the Reader Service Number below and fax this page to +1 (416) 620-9790. Include your name, company, complete
H 䊐-Service Company
address, phone number, fax number and e-mail address, and check the box on the right for your division of industry
J 䊐-Chemical Co.
and job title.
P 䊐-Engrg./Construction Co.
Name ________________________________________________________ Company ________________________________________________________ JOB FUNCTION
Address ______________________________________________________ City/State/Zip ____________________________________________________ (check one only):
B 䊐-Company Official, Manager
Country ______________________________________________________ Phone No. _______________________________________________________
E 䊐-Engineer or Consultant
FAX No. ______________________________________________________ e-mail ___________________________________________________________ F 䊐-Supt. or Asst.
G 䊐-Foreman or Asst.
This Advertisers’ Index and procedure for securing additional information is provided as a service to Hydrocarbon Processing I 䊐-Chemist
advertisers and a convenience to our readers. Gulf Publishing Co. is not responsible for omissions or errors. J 䊐-Purchasing Agt.
ADVERTISERS in this issue of HYDROCARBON PROCESSING

Company Page RS# Company Page RS# Company Page RS#
Website Website Website
ABB Inc. . . . . . . . . . . . . . . . . . 18 (73) ExxonMobil . . . . . . . . . . . . . . . 20 (77) M3 Technology . . . . . . . . . . . . 35 (154)

www.info.hotims.com/23161-73 www.info.hotims.com/23161-77 www.info.hotims.com/23161-154
Ametek Process Instruments . . . 6 (123) Flexitallic LP . . . . . . . . . . . . . . . 5 (93) MSA Instrument Division . . . . . 63

www.info.hotims.com/23161-123 www.info.hotims.com/23161-93 NPRA . . . . . . . . . . . . . . . . . . . 87 (54)
Armstrong International Inc . . . 36 (68) Gulf Publishing Company www.info.hotims.com/23161-54
www.info.hotims.com/23161-68 GPC Books . . . . . . . . . . . . . . 84 (163) Paharpur Cooling Towers, Ltd. . 27 (57)

Armstrong International Inc . . . 37 (155) www.info.hotims.com/23161-163 www.info.hotims.com/23161-57
www.info.hotims.com/23161-155 GPC Books . . . . . . . . . . . . . . 43 (157) Paratherm . . . . . . . . . . . . . . . . 64 (160)

Axens . . . . . . . . . . . . . . . . . . . 88 (53) www.info.hotims.com/23161-157 www.info.hotims.com/23161-160
www.info.hotims.com/23161-53 GPC Books . . . . . . . . . . . . . . 68 (162) Rentech Boiler System . . . . . . . . 2 (64)
BASF Catalysts LLC . . . . . . . . . 16 (83) www.info.hotims.com/23161-162 www.info.hotims.com/23161-64
www.info.hotims.com/23161-83 Circulation . . . . . . . . . . . . . . 76 Samson GmbH . . . . . . . . . . . . . 4 (151)
Burckhardt Compression AG . . 38 (55) Software Reference Guide www.info.hotims.com/23161-151
www.info.hotims.com/23161-55 Showcase . . . . . . . . . . . . . 75
SNC-Lavalin Engineers &
HPI Classified Pages . . . . . 82–83
Coade Engineering Software . . 23 (152) Construction Inc.. . . . . . . . . . 26 (153)
www.info.hotims.com/23161-152
John M Campbell & Co . . . . . . 65 (161)
Cudd Energy Services . . . . . . . 47 (158) Spraying Systems Co . . . . . . . . 14 (62)
KBC Advanced
Technologies Inc . . . . . . . . . . 28 (82)
Emerson Process Taper-Lok . . . . . . . . . . . . . . . . 42 (156)
Management (DeltaV) . . . . . . 12 (65)
www.info.hotims.com/23161-65 KTI Corporation . . . . . . . . . . . . 44 (96)
www.info.hotims.com/23161-96 Thomas Russell Co . . . . . . . . . 10 (74)
Emerson Process
Management (Fisher) . . . . . . 22 (69) Linde Process Plants . . . . . . . . 24 (79)
www.info.hotims.com/23161-69 www.info.hotims.com/23161-79 UOP LLC . . . . . . . . . . . . . . . . . . 8
For information about subscribing to HYDROCARBON PROCESSING, please visit www.HydrocarbonProcessing.com 85
HPIN CONTROL
Y. ZAK FRIEDMAN, CONTRIBUTING EDITOR
[email protected]
CDU overhead double-drum configuration

Fig. 1 shows one of many CDU overhead configurations, naphtha, and mixing it into the product creates an undesirable
with naphtha cutpoint control accomplished by a column top heavy tail. Ten percent of the reflux is light kero material, and
temperature controller manipulating a top pumparound (TPA) downgrading kero to reformer feed in today’s prices carries a
circuit heat removal. This configuration is heat efficient although penalty of about $7 per bbl. Even if there is a price reversal, good
heat efficiency comes at the expense of top section separation. The separation between kero and naphtha would be profitable, and
TPA internally uses four trays, not for separation, but for internal that is why the double drum is there in the first place.
condensing and heat transfer. Another feature I dislike about Fig. 2 is the method of inferring
To regain those lost trays process designers often specify two naphtha cutpoint. The temperature most indicative of naphtha
overhead drums per the configuration of Fig. 2. Heat previ- cutpoint is not the TC on top of the column but rather the blue
ously removed by the TPA circuit is now removed by a reflux TI on the reflux drum.
condenser against crude, more or less at the same temperature What I consider a thermodynamically correct way of control-
levels. Vapor from the reflux drum is further condensed into a ling a double-drum overhead system is illustrated in Fig. 3. The
product drum. On the whole—a thermodynamic system that combination of reflux drum level control on the reflux, and naphtha
not only gives us more top section trays, but the reflux drum is cutpoint control on the reflux condenser elminates excess reflux. For
also a separation stage. good dynamic response, tune the blue level controller tightly. IE,
Examine now the DCS control of Fig. 2. Naphtha cutpoint apply a strong controller gain, but beware of making the reset action
is controlled by manipulating reflux instead of TPA duty. Excess too aggressive and driving the controller unstable.
reflux drum material is blended into the naphtha product. Is it a Fig. 3 permits recycle of product naphtha into the reflux drum
good idea to mix reflux into the product? Reflux is heavier than but that is used only in abnormal situations. Such recycle may
become necessary during hot summer hours, when even maximum
reflux condenser operation cannot maintain the naphtha cutpoint
Spillback at target. Bear in mind that the recycle of naphtha into the reflux
drum is in the category of reflux going down the column and is
not thermodynamically damaging, but it is not desirable because it
PC replaces high-temperature cooling against crude by low-temperature
Offgas
cooling against air. HP
TC
FC
LC
The author is a principal consultant in advanced process control and online
Crude optimization with Petrocontrol. He specializes in the use of first-principles models
TPA FC for inferential process control and has developed a number of distillation and reactor
Naphtha models. Dr. Friedman’s experience spans over 30 years in the hydrocarbon industry,
working with Exxon Research and Engineering, KBC Advanced Technology and since
FIG. 1 Single-drum overhead. 1992 with Petrocontrol. He holds a BS degree from the Israel Institute of Technology
(Technion) and a PhD degree from Purdue University.
Spillback Spillback
Crude Crude
TI TC
PC TC
Offgas Offgas
Normally
closed
LC LC LC LC
TC TI
FC FI FC FC FI FC
Naphtha Naphtha
FIG. 2 Double-drum overhead. FIG. 3 Ideal double-drum control.
86
DFH79ED<;H;D9;I
?\OekÊh[Dej>[h["OekÊh[@kijEkjJ^[h[$
DFH7H[b_WX_b_joCW_dj[dWdY[9ed\[h[dY[WdZ;n^_X_j_ed
CWo'/Å(("(&&/¼=WobehZJ[nWdH[iehj9edl[dj_ed9[dj[h¼=hWf[l_d["JN
:edÊj][jbeYa[Zekje\DFH7ÊiH[b_WX_b_joCW_dj[dWdY[9ed\[h[dY[$
?djeZWoÊiZ[cWdZ_d][Yedeco"oekmWdjceh[lWbk[\ehoekhZebbWh$DFH7ÊiH[b_WX_b_jo
CW_dj[dWdY[9ed\[h[dY[WdZ;n^_X_j_edZ[b_l[hi_jm_j^Wd[nfWdZ[Zfhe]hWcj^Wj_dYbkZ[i
ceh[j^Wd*&mehaiêfi_di[l[dikX`[YjjhWYai"jmea[odej[if[Wa[hi"WdZi[l[hWb
G7WdZZ_iYkii_edfWd[bi$
J^[;n^_X_j_edm_bbiêmYWi[j^[bWj[ijj[Y^debe]_[iWdZi[hl_Y[i\ehoekjei[[WdZ[lWbkWj[$
Ceh[j^Wd(*&YecfWd_[im_bb[n^_X_jj^[_h[gk_fc[dj"cWj[h_Wbi"WdZif[Y_Wb[nf[hj_i[_d
iebl_d]fheXb[cij^WjWh[Yh_j_YWbjeoekhYecf[j_j_l[d[ii$
J^[(&&/HC9_ioekhX[ijeffehjkd_jojei[[WdZ^[WhWXekjj^[bWj[ij_dh[b_WX_b_jo"
cW_dj[dWdY["WdZjkhdWhekdZi\hecdej[Z[nf[hjiWdZ\[bbemfhWYj_j_ed[hi$
H[]_ij[hjeZWoWjdfhW$eh]%hcYeh
YWbbkiWj(&($*+-$&*.&\ehceh[_d\ehcWj_ed$
NATIONAL PETROCHEMICAL JeZWoÊih[Ód_d]WdZf[jheY^[c_YWbXki_d[ii[i$

& REFINERS ASSOCIATION J^[ceh[oekademWXekjki"j^[X[jj[hm[beea$
Our advanced
catalytic engineering
has powerful attraction
Axens’ HR series ACE™ technology catalysts deliver sustained, on-spec ultra-
low-sulfur diesel. The powerful desulfurization activity of this new series is
particularly enhanced in combination with EquiFlow™ reactor internals and
Catapac™ dense loading technology.
With its submicron-level control, ACE technology provides dual activity for
superior levels of sulfur and nitrogen removal in all applications. Easy to load
and activate, HR series catalysts also offer excellent regeneration properties.
Axens’ quality products are based on reference technology – commercially

proven, dependable, reliable and cost-effective.
The performance improvement specialists

www.axens.net
For more information
Paris +33 1 47 14 25 14 Houston +1 713 840 1133 Email [email protected]

Hydrocarbon Processing 2019-03

Uploaded by

Copyright:

Available Formats

Hydrocarbon Processing 2019-03

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hydrocarbon Processing 2019-03

Uploaded by

Copyright:

Available Formats

What are some of the topics covered in the special report on instruments and networks?

What are some of the topics covered in the special report on instruments and networks?

What are some challenges of constructing liquefaction facilities in the Arctic Circle?

What are some challenges of constructing liquefaction facilities in the Arctic Circle?

HYDROCARBON PROCESSING

HPIMPACT SPECIALREPORT BONUSREPORT

Top initiatives INSTRUMENTS GAS PROCESSING

Nobel Laureate Wireless, soft sensors, New methods

Fired Package Boilers / Wasteheat Boilers / Heat Recovery Steam Generators

SPECIAL REPORT: INSTRUMENTS AND NETWORKS

29 Wireless networks improve refinery operation

39 Soft sensor modeling using artificial neural networks

45 Hydrogen gas detection

55 What are the opportunities to construct liquefaction

66 Auxiliary pumps and support systems

PROCESS DEVELOPMENTS COLUMNS

DEPARTMENTS 86 HPIN CONTROL

GULF PUBLISHING COMPANY

ALSO AVAILABLE IN:

New AMETEK 5100 NCM™ laser analyzer

HYDROCARBON PROCESSING MARCH 2009

Our depth of knowledge and experience gives UOP

As the global leader in technology solutions for the petroleum

Process Technology • Catalysts • Adsorbents • Performance Equipment • Profitability Consulting

Unreliability, global procurement and you

DESIGN • BUILD • INSTALL

7050 South Yale, Ste. 210

Sacrificed to the money system: engineering workforce

A good alarm management strategy

Spray Injector Solutions

Learn More at spray.com/injectors

Our solutions include injectors for: Computational Fluid Dynamics

In the US and Canada: 1-800-95-SPRAY | 1-630-665-5000 | spray.com | [email protected]

Association news in brief

HYDROCARBON PROCESSING MARCH 2009

Trust BASF Reﬁning Catalysts

Source: ISA, In Tech, January 2009

HYDROCARBON PROCESSING MARCH 2009

Unrivalled experience you can trust.

• Delayed Coker Process Engineer - B.S. Chemical or

Please apply online at exxonmobil.com/ex to the job numbers listed above.

Exxon Mobil Corporation An Equal Opportunity Employer

HYDROCARBON PROCESSING MARCH 2009

Select 152 at www.HydrocarbonProcessing.com/RS

OpX – Energy Efficiency Initiative, European Refiner.

Wireless networks improve

key components of TSMP that contribute

Innovative wireless applications grow at BP

security policies to allow only trusted end-point devices to access

Wireless field networks. A variety of wireless field devices

voice, mobility and tracking. The wireless plant network is illus-

OPC UA: an end user’s perspective

Cut field connections

Security: the new challenge for automation. OPC

Randy Kondor is a computer engineer and president of the

Select 155 at www.HydrocarbonProcessing.com/RS

For longest running time: We recommend our own designed,

Designed for easy maintenance

We are the competent partner with the full range

More beneﬁts: www.recip.com/api618